The region to be discussed comprises the two Iranian provinces of West Azerbaijan and East Azerbaijan, with administrative centers at Urmia (before 1979 Reżāʾīya) and Tabrīz respectively; it does not include “Northern Azerbaijan,” centered on Baku, which since 1829 has belonged to the Russian empire.
The modern provincial and international boundaries do not correspond to limits of ethnic or tribal areas. The border between the provinces of Kurdistan and West Azerbaijan and the frontier on the Aras (Araxes) river between East Azerbaijan and Russian Azerbaijan cut across such areas. Before the partition in the nineteenth century, Iranian and Russian Azerbaijan constituted a single cultural entity. In ancient times, however, the two provinces now belonging to Iran had formed a distinct cultural region, known as Media Atropatene because, after the collapse of the Achaemenid empire, the satrap Atropates secured the political independence of this part of the former satrapy of Media (Pauly-Wissowa, II, col. 2150).
Azerbaijan is a mountainous region where routes of ancient origin intersect. It has thus been, throughout the centuries, both a pole of attraction for migrating peoples and warring armies and a center of commercial and cultural exchange. It was the bridge from Mesopotamia to the metal-rich lands of the Caucasus and from the Anatolian plateau to central Iran, with further links to Transcaucasia and India.
Ancient sites. From the pre-historic period onward, Azerbaijan was at least sparsely populated. The oldest known traces of human settlement are Paleolithic cave-dwellings, such as the cave at Tamtama, north of Urmia in West Azerbaijan, found by C. Coon (Cave Explorations in Iran 1949, Philadelphia, 1951, pp. 15-20), and the caves which, together with some open-air sites, have been found in the Sahand massif south of Tabrīz in East Azerbaijan (survey report in Iran 14, 1976, p. 154). It was apparently not until the late Neolithic period, from 6,000 B.C. onward, that Azerbaijan came under closer human occupation. Evidence of this has been brought to light by the British excavations at Yanīk Tepe on the east shore of Lake Urmia (C. A. Burney, “Excavations at Yanik Tepe, North-West-Iran,” Iraq 23, 1961, pp. 138ff.) and by the findings of the American Ḥasanlū project in the Soldūz plain and at Ḥasanlū itself (see R. H. Dyson, “Hasanlu 1974. The Ninth Century B.C. Gateway,” in Proceedings of the 3rd Annual Symposium on Archaeological Research in Iran 1974, Tehran, 1975, pp. 179ff.; I. E. Reade, “Hasanlu, Gilzanu and Related Considerations,” AMI 12, 1979, pp. 175ff.; Vanden Berghe, Bibliographie analytique, pp. 157ff., and Supplément 1, pp. 46ff.). Pottery from then on shows vigorous development of both shape and decoration. Azerbaijan in the phase of incipient continuous settlement offers one of the Near East’s most interesting fields for archeological exploration, as can be seen from the results of the fruitful efforts made mainly between the end of the Second World War and the Islamic revolution (ca. 1950-78). Surveys and test diggings have produced evidence of close settlement in different periods around Lake Urmia and of inhabited sites and forts in valleys leading up into the mountains.
The population of the west bank of Lake Urmia became denser from the fourth millennium B.C. onward. Surveys conducted by Italian archeologists in an area west of Urmia (P. E. Pecorella and M. Salvini, Fra lo Zagros e l’Urmia: Ricerche storiche ed archeologiche nell’Azerbaigian Iraniano, Rome, 1984), by the German Archeological Institute in the northwest of Azerbaijan (Vanden Berghe, op. cit., nos. 1964-70, 1973, 2240-54), and by several smaller expeditions have revealed numerous sites of settlements of the third and second millennia in Azerbaijan. As yet it has been possible only to take measurements, but not to start excavations, at Ravaz, a relatively large settlement north of Sīah Čašma, and at Yaḵvalī, a fort settlement east of Mākū (W. Kleiss and S. Kroll, AMI 12, 1979, pp. 27ff.). Each is from the third millennium, having round houses of a type already known from the excavations of the third millennium sites at Haftavān Tepe and Yanīk Tepe. Both belong to the Early Caucasian culture (early Bronze Age).
Ravaz presents a vivid picture of a big and important settlement in the third millennium. It was defended by a thick stone wall, later supplemented with stout semicircular towers in front. Access to the houses was through a single tongue-shaped gateway. The houses were packed tightly together, but traces of streets are discernible. This site was never built over in later times. Outside the main settlement lay an extensive periphery of seemingly terraced fields or gardens with single round houses. Lines of roads can be recognized in this area also.
Yaḵvalī, on the other hand, is a small fort settlement. It too had a solid defensive wall, but without any towers. The access gateway was of simple design. Outside the fort lay some separate groups of round houses.
Bolūrābād, northeast of Besṭām, is another fort settlement from the third millennium B.C. (idem, AMI, N.S. 8, 1975, pp. 15ff.). Its surrounding wall was 3 m thick and of quarried stone. Remains of round houses can be seen inside. In a second phase of building, probably still in the third millennium, the wall was strengthened at its most vulnerable point by the addition of an external box-like structure which was filled with earth. At this site also, only measurement has so far been possible.
In the northwestern part of the province of West Azerbaijan, measurements of extensive tumulus clusters at Maḵand, Qara Żīāʾ-al-dīn/Besṭām, and Maryam northeast and east of Mākū, and of a single tumulus at Vār west of Ḵoy, have been carried out (Kleiss, ibid., 11, 1978, pp. 13ff.). Although none of these tumuli have yet been opened, there can be no doubt that they date from a period extending from the second into the first millennium B.C., as do the graveyard sites in northeastern Azerbaijan between Meškīnšahr and Ardabīl and in the Ṭāleš mountains on the Caspian west coast. The abundance of tumuli in the pastureland of the Aras plain between Mākū and Besṭām indicates that they were graves of the equestrian nomadic peoples who roamed in this area before the arrival of the Urartians ca. 800 B.C. For this reason no remains of settlements connected with the tumuli are likely to be found. Different methods of tumulus erection have been noted, those at Maḵand being made of earth with stone trimmings (the commonest type), those at Maryam made of earth and stones, and those at Qara Żīāʾ-al-dīn/Besṭām made entirely of stones. Finds of importance for knowledge of the general development of architecture, ceramics, and burial practices have emerged from the excavation of Kordlar Tepe by Austrian archeologists (A. Lippert, “Die Österreichischen Ausgrabungen am Kordlar-Tepe in Persisch-Westaserbeidschan (1971-78),” AMI 12, 1979, pp. 103ff.). Most of the finds are from the early Iron Age (11th century B.C.). This settlement had a central building of fort-like design.
Haftavtān Tepe, which was explored by British archeologists, is potentially one of the more important sites in the northwest of Iran, having been inhabited from the fourth millennium through the Urartian period right down to Sasanian times (C. A. Burney, “The Fifth Season of Excavations at Haftvan Tappeh: Brief Summary of Principal Results,” in Proceedings of the 4th Annual Symposium on Archaeological Research in Iran 1975, Tehran, 1976, pp. 257ff.). The British excavations at Goy (Gök) Tepe south of Urmia have yielded interesting evidence of cultural links between the plain on the west bank of Lake Urmia, Mesopotamia, Anatolia, and Central Iran in the second and first millennia B.C. (T. Burton-Brown, “Geoy Tepe,” in Excavations in Iran, the British Contribution, Oxford, 1972, pp. 9-10). Also noteworthy are the investigations done by J. and H. de Morgan in cemeteries of the third to second millennia in the Ṭāleš district in the northeast of Azerbaijan (H. de Morgan, “Recherches au Talyche persan,” in MDAP VIII, 1905, pp. 251ff.).
The biggest and richest prehistoric burial site in Azerbaijan, however, is the one at Ḥasanlū. The adjoining settlement, already occupied from the third to the sixth millennium, was strengthened in a later period (Ḥasanlū IV) with a citadel and annexes. These so-called “burned buildings” mark a significant step in the evolution of large-room construction from the Hittite architecture of the fortress at Boğazköy to the later Urartian architecture and thence to the Median halls at Godīn Tepe west of Hamadān (T. Cuyler Young, “The Chronology of the Late Third and Second Millennium in Central Western Iran as Seen from Godin Tepe,” American Journal of Archaeology 73, 1969, pp. 287ff.) and the Achaemenid apadānas at Persepolis and Susa (E. F. Schmidt, Persepolis I, Chicago, 1953). The layout of the Ḥasanlū IV citadel foreshadows the fortress architecture of the first millennium B.C., and the rich finds of metallic and ceramic objects are typical of its early centuries (R. H. Dyson, “Architecture of the Iron I Period at Hasanlu in Western Iran and Its Implications for Theories of Migration on the Iranian Plateau,” in Le Plateau iranien et l’Asie Centrale, Paris, 1977, pp. 155ff.). Ḥasanlū was probably conquered, plundered, and destroyed by the Urartians ca. 800 B.C., but some time later it was rebuilt as a strong Urartian fortress.
Urartian period. From ca. 800 to the mid-7th century B.C., the Urartians held the districts southwest, west, northwest, and northeast of Lake Urmia. Thus the whole of the modern province of west Azerbaijan except the southern district around Mīāndoāb, and the western part of East Azerbaijan up to somewhere near Ahar, belonged to Urartu. By 1978 a total of 101 Urartian forts, settlements and other sites, and inscriptions had been identified, including six inscriptions on rocks and buildings already known before 1967 (Kleiss and Kroll “Vermessene urartäische Plätze in Iran (West-Azerbaidjan) und Neufunde (Stand der Forschung 1978),” AMI 12, 1979, pp. 183ff.). The principal excavations were done at Besṭām/Rusa-i URU.TUR (Vanden Berghe, op. cit., nos. 2217-40 and 4278-300; on Sangar see also Istanbuler Mitteilungen 18, 1968, pp. lf.), Haftavān Tepe (Ch. Burney, “Excavations at Haftavan Tepe 1969,” Iran 10, 1972, pp. 127ff.), Qaḷʿa-ye Esmāʿīl Āqā (Pecorella and Salvini, Fra lo Zagros e l’Urmia, pp. 215ff.), Ḥasanlū and Agrab Tepe (O. W. Muscarella, “Excavations at Agrab Tepe, Iran,” The Metropolitan Museum Journal 8, 1973, pp. 47ff.), and Moḥammadābād southwest of Urmia (the Iranian excavations are not yet published; W. Kleiss, AMI, N.S. 9, 1976, pp. 36ff.; the inscription has been published by M. Salvini, AMI, N.S. 10, 1977, pp. 125ff. and in Pecorella and Salvini, op. cit., pp. 77f.).
Besṭām was founded by the Urartian king Rusa II (685-645 B.C.), who has left an inscription on stone from a building. According to the inscription, which has been moved from Besṭām to the Mūza-ye Īrān-e Bāstān at Tehran, the name of the newly founded settlement was Rusa-i URU.TUR “Rusaδs town.” Here a brief description of the Urartian site at Besṭām will suffice (see BESṬĀM).
Besṭām consists of a citadel, a craftsmen’s and tradesmen’s quarter, and a square walled enclosure at the foot of the citadel hill probably used for keeping horses. The citadel comprised a lower part, reserved mainly for the garrison but also containing stables, business premises, and a guest house; a middle part, stretching up the slope of the hill, with large storage areas for supplies needed in the fortress, and on the upper levels ceremonial halls and the temple of the Urartian god Haldi (which is mentioned in the inscription); and a highest part or acropolis which was used as a royal lodging and as a last refuge in emergencies. Besṭām/Rusa-i URU.TUR appears to have been the regional base from which the itinerant king or his governors controlled Urartu’s eastern territories. The citadel at Besṭām is the biggest-known building complex of the Urartians, covering a larger area than the citadels of their capitals at Van and Toprak Kale in Turkey. The most recent excavation in 1978 produced evidence that Besṭām was probably plundered and burned during civil wars in the second half of the seventh century B.C.
No Urartian cemeteries have been found at or near Besṭām, but some Urartian rock-chamber tombs in Azerbaijan are known. The most impressive is a series of three chambers in the vicinity of a rather small Urartian fort at Sangar, west of Mākū (Kleiss, Istanbuler Mitteilungen 18, 1968, pp. 1ff.). Access is by a 1.10 m-wide staircase of thirty steps cut in the rock face. The tomb-chambers consist of a main chamber, 3.15 m in height and 5.45 x 3.90 m2 in area, with niches in the walls and a big niche-like recess, 2.00 m in height and 2.30 x 1.30 m2 in area, and of two side chambers reachable from the main chamber through doorways. The dimensions of the recess at the back of the main chamber are sufficient to take a sarcophagus. In the middle of the rear wall of the recess a slab-shaped alcove has been chiseled out, perhaps for the placing of a stela.
The Urartian sites are sufficiently numerous to permit the drawing of a rough map of the Urartian road network. The roads connected the various places to each other and led westward to the heartland of Urartu around the capital Tuspa (Van in Turkey). It has also been possible to locate a number of staging posts, some close to passes. One of these, at Tepe Dosoğ near the pass between Urmia and Ošanavīya, prefigures the courtyard type of hostelry which, in its eventual development, was to become important as the normal form of the oriental caravansary (Kleiss and Kroll, AMI 12, 1979, pp. 195f.). Still to be seen near the fort and settlement at Veraḵram are vestiges of an Urartian bridge over the Aras (now the river frontier with U.S.S.R.); this is the oldest bridge known to have existed in Azerbaijan (ibid., p. 221).
Somewhat antedating or perhaps contemporary with the Urartian period in Azerbaijan is the probably Mannean sanctuary at Zendān-e Solaymān, which lay beyond Urartu’s southern limits. The German excavations at this site have shown that it had an unusual layout, with a surrounding wall in the form of a series of box-like structures not yet seen elsewhere in the region. The wall surrounded a natural crater-lake formed by sinter deposits. Construction took place mainly in two phases. The sanctuary, a terraced structure, was built first, in the eighth century B.C. Later the sanctuary was abandoned and the site was walled and fortified to serve as a safe haven for the Manneans (Kleiss, Zendan-i Suleiman: Die Bauwerke, Wiesbaden, 1971). The artifacts found at Zendān-e Solaymān resemble those found at Zīvīya (Ziwiyeh) in the nearby province of Kurdistan (A. Godard, Le Trésor de Ziwiye (Kurdistan), Haarlem, 1950, p. 136).
Armenian monuments. In connection with the Armenians, certain forts in the area north to northeast of Mākū deserve mention. Because of the architectural features of the remains and the pottery found in them, they must be dated from the sixth century B.C. It has been possible to take measurements of two of them Īlān Qara II and Qaḷʿa-ye Ḥājjestān (despite the superimposition of a medieval Armenian castle on the latter site). They have the salient-reentrant wall lines typical of Azerbaijan in the sixth century B.C. Both must therefore be classified as purely Urartian forts. But since the territory and culture of the Urartians were taken over by the Armenians, sites such as Īlān Qara II and Qaḷʿa-ye Ḥājjestān are likely to have become early (pre-Christian) Armenian forts. Elsewhere in West Azerbaijan there are remains of other forts of prehistoric origin which remained in use until the later Middle Ages. Some names of Armenian castles in the Armenian province of Vaspurakan are known from old writings but difficult to pin on ruins visible today. For example, it has not yet been possible to find the name of the Armenian castle which stood on top of the Urartian ruins at Besṭām from the ninth to the fifteenth century A.D. (see also further below).
Median period. It is not yet possible, in the present state of knowledge, to describe the political changes which led to the ending of Urartian rule in Azerbaijan in the middle or the second half of the 7th century B.C. Some clues, however, have been provided by archeological investigations in the northwestern part of West Azerbaijan. A change in the method of fortress defense is indicated by the construction of massive walls, with the rectilinear, salient, and reentrant lines of ramparts, on the irregular contours of citadel hills. The pottery also shows change, with the development of the painted, so-called “triangle ware.” In many cases the dating of architectural and ceramic remains in the transition from the Urartian to the Median and subsequent Achaemenid periods has not yet been satisfactorily worked out. There is reason to believe that certain remains of settlements at Besṭām and in its vicinity are Median. Investigations have shown, however, that the rock-tomb at Faḵraka south of Lake Urmia, which was once thought to be Median, is of considerably more recent, probably late Achaemenid, origin (H. von Gall, “Zu den "Medischen" Felsgräbern in Nordwestiran and iraqi Kurdestan,” Archäologischer Anzeiger, 1966, pp. 20ff.).
Achaemenid period. Azerbaijan was annexed to the empire of the Achaemenids some time in the second half of the sixth century B.C. They have left very few relics in this province compared with others. Some graves and houses of the period have been discovered at Taḵt-e Solaymān and other cemetery sites (R. and E. Naumann, “Takht-i Suleiman,” in Katalog der Ausstellung München 1976, p. 26). Finds of Achaemenid pottery have been made at Qaḷʿa-ye Żaḥḥāk (Kleiss, AMI, N.S. 6, 1973, pp. 163ff.)
Seleucid and Parthian periods. The Seleucids (312-129 B.C.) have left no significant vestiges in Azerbaijan. From the Parthian period (191 B.C.-A.D. 225), however, remains of settlements and cemeteries are widely scattered throughout the region. Numerous graveyards of the first century A.D. have been found in the Ṭāleš district around Germī north of Ardabīl (Iranian excavations directed by S. Kāmbaḵš-e Fard, unpublished report). At Taḵt-e Solaymān, however, the excavations have yielded so little in the way of Parthian pottery that the identification of this site with the Parthian fortress Phraaspa, which the Roman general Mark Antony besieged without success in 36 B.C., is no longer tenable (R. and E. Naumann, op. cit., p. 11). Phraaspa is more likely to be traceable in the Marāḡa district. At Qaḷʿa-ye Żaḥḥāk south of Sīāh Čaman (Qara Čaman) in East Azerbaijan, extensive ruins of buildings from the Parthian period, including an almost wholly intact brick-walled pavilion of the first century A.D., await detailed examination (Kleiss, AMI, N.S. 6, 1973, pp. 163ff.). The design of the pavilion’s facade exhibits a blend of old Iranian traditions of building inherited from Achaemenid times with influences from Roman architecture. There are strong grounds for the hypothesis, first propounded by V. Minorsky (BSOAS 9, 1943-46, p. 262) that Qaḷʿa-ye Żaḥḥāk is to be identified with the Parthian town Phanaspa mentioned by Ptolemy (Geography, ed. F. W. Wilberg, VI, 2, Essendiae, 1838-45, p. 393).
Sasanian period. The Sasanian period (A.D. 240-642) is represented in Azerbaijan by clearly identifiable remains of settlements, including one at Besṭām, by parts of mosques, such as the Masjed-e Jomʿa at Urmia, and by ruins of fortresses. Many old castles are believed to be of Sasanian origin, and in some of them Sasanian brick or stone work can be recognized, but more often the belief rests on speculation rather than factual evidence of Sasanian characteristics of the walls. No solution has yet been found to the problem of distinguishing Sasanian from Islamic fortification techniques. The same is true of pottery designs.
The most important legacy of Sasanian art in Azerbaijan is the rock-carving at Salmās (formerly Šāhpūr). This dates from the third century A.D. and probably represents an act of homage or acceptance of vassalage by the Armenians in the presence of Ardašīr I and the crown prince Šāpūr. The lowness of the relief sets it apart from the more sculptural Sasanian rock-carvings in other regions (W. Hinz, “Das Sasanidische Felsrelief von Salmās,” Iranica Antiqua 5, 1965, pp. 148ff.). A Sasanian rock-inscription survives in the Meškīnšahr district in East Azerbaijan (ibid., nos. 2266-67).
The German excavations at Taḵt-e Solaymān, the site of the Sasanian sanctuary of Ādur Gušnasp, have thrown light on one of the most important and interesting cult centers of the Zoroastrian religion (R. Naumann, “Die Ruinen von Tacht-e Soleiman und Zendan-e Suleiman und Umgebung,” in Führer zu archäologischen Plätzen in Iran II, Berlin, 1977. D. Huff, “Recherches archéologiques à Takht-i Suleiman, centre religieux royal sassanide,” Comptes rendus de l’Académie des inscriptions et belles-lettres, 1978, pp. 774ff.). The fire sanctuary, known as the sanctuary of Šīz, had originally stood at Ganzak (probably identifiable with the Laylān) and was apparently moved to the plateau now called Taḵt-e Solaymān (Solomon’s throne) by Ḵosrow I in the middle of the sixth century A.D. The fire temple, which was the focal point of the complex, evidently suffered at least partial destruction at the hands of the Byzantine army of the emperor Heraclius in 624, although according to Abū Dolaf Mesʿar b. Mohalhel, who wrote in the mid-4th/10th century, the fire had been burning for seven hundred years and was still alight in his time. The excavations have produced evidence that the fire sanctuary buildings which finally took shape on the Taḵt-e Solaymān plateau remained intact for only about one hundred years, from Ḵosrow I’s to Ḵosrow II’s reign. Before then, perhaps in the fifth century during the reign of Pērōz I, the plateau had been fortified for the first time with the construction of a mud-brick wall, and mud-brick buildings had been erected inside the sacred enclosure. The plan devised at that time was on a large scale, and the builders in the later Sasanian period adhered to its general lines. The plateau is a terrace of sinter built up by lime deposition from a powerful spring, the impounded waters of which form a small lake or pond in the middle of the terrace. The original mud-brick buildings and surrounding wall of the sanctuary were replaced from ca. 500 onward by structures of hewn stone with baked brick vaults. Sasanian building activity at Taḵt-e Solaymān certainly reached its peak in the reign of Ḵosrow I (531-79), when the Ādur Gušnasp fire was relocated and suitably housed at this site, then called Šīz. After the destruction by the Byzantines and the subsequent conquest of Iran by the Arabs, practice of the fire cult at Taḵt-e Solaymān continued for a long time, despite the establishment and gradual expansion of a Muslim settlement on the plateau in the ʿAbbasid period. Even so, Taḵt-e Solaymān lost the religious eminence which it had enjoyed under the Sasanians. It did not regain any sort of importance until ca. 1271, when the il-khan Abaqa (Abāqā) built a summer palace on top of the ruins, partly incorporating walls and surviving chambers of the former fire sanctuary.
Islamic period. Azerbaijan, with the rest of Iran, fell to the Muslims in the mid-seventh century A.D. Its history in the first four centuries of the Islamic period is to a large extent obscure. The oldest surviving Islamic edifices were built in the Saljuq period, e.g., the Se Gonbad tomb-tower at Urmia which dates from 1180 and is notable for the fine stalactite stucco ornamentation over its portal. The domed hall of the great mosque (Masjed-e Jomʿa) at Urmia is conceptually derived from the Sasanian čahār-ṭāq (dome on four arches over a fire altar) and evidently stands on pre-Islamic foundations. This mosque’s meḥrāb dating from 1277 and another at Marand dating from the fourteenth century are fine examples of the use of stucco for niche-decoration in the Il-khanid period, comparable in the delicacy of their carving with the meḥrāb in the Masjed-e Jomʿa of Isfahan. Also important are the remains of the Masjed-e Jomʿa and its minaret at Ardabīl from the Saljuq or the Il-khanid period (12th or 13th century). The tomb-tower at Meškīnšahr from the Il-khanid period or the time of Tīmūr (13th or 14th century) is the last big tomb-tower left in Azerbaijan since the destruction of the one at Salmās in an earthquake in 1930. Several smaller tomb-towers, mostly of later date, also exist and have been surveyed. (On the Islamic architecture in general, see Survey of Persian Art I-IX, London 1938; XIV, New York and Tehran 1967; XV, Tehran, 1977. For the survey reports, see Kleiss in AMI, N.S. 26, 1969-73.) At Tabrīz, nothing from the early Islamic period remains, and the Masjed-e Jomʿa in the bazaar, though originally built in the Saljuq period, underwent drastic alteration in the fifteenth century. Of the five architecturally important tomb-towers at Marāḡa, the Qoy-borj (Tower of the ram) from the Timurid period has collapsed; three of those still standing are from the Saljuq period, namely the Gonbad-e Sorḵ (Red dome) completed in 1148, which is the oldest, the Gonbad-e Kabūd (Blue dome), and a round tower, while the fourth, known as the Gonbad-e Ḡaffārīya, is from the time of the Il-khans (early 14th century).
As already mentioned, the il-khan Abaqa (1265-81), whose capital was at Marāḡa, had a hunting lodge, named Saturiq, built on the Taḵt-e Solaymān plateau. The walls of its main rooms were richly adorned with carved stucco and glazed tiles (R. and E. Naumann, “Takht-i Suleiman,” pp. 43ff., 61ff.). Being built on top of the Sasanian ruins, the Mongol palace on the whole conformed to the plan of the Zoroastrian sanctuary with the pond at the center of the layout, but the main entrance was shifted from the north to the south side of the defensive wall around the plateau. Pillared galleries were built around the pond on all four sides, probably also as in Sasanian times. Behind the arcades lay rooms with differing but commodious dimensions. On the site of the former fire-temple, a large and conspicuous chamber with a north-south orientation was erected, possibly for use as an audience hall, and made accessible by means of a flight of steps. Among the decorative objects found in the Il-khanid palace at Taḵt-e Solaymān, a marble capital with the acanthus design is particularly interesting; it is typical of the late Roman period and likely to have been imported from northern Syria. Remains of kilns and potters’ workshops at Taḵt-e Solaymān indicate that the wall tiles were manufactured on the spot. Taḵt-e Solaymān was abandoned early in the fourteenth century and Abaqa Khan’s palace then fell into ruin (ibid., p. 12).
As regards the dating of the foundation of Azerbaijan’s main towns, no clear evidence is available. Despite the lack of definite proof, it can be taken for certain that Urmia is of pre-Islamic origin. There are no archeological remains to show that Tabrīz existed in pre-Islamic times, and the earliest date of its establishment given in a literary source is 175/791 during the caliphate of Hārūn al-Rašīd (Nozhat al-qolūb, p. 75). In general the larger towns of Azerbaijan appear to have come into being soon after the spread of Islam in Iran, because most of them contain buildings or remains of the early Islamic period. Tabrīz was the capital of the il-khans in the 7th/13th century and of the Qara Qoyunlū and Āq Qoyunlū Turkman dynasties in the 9th/15th century. Surviving from the Turkman period are the remains of the Masjed-e Kabūd (Blue Mosque) and from the Il-khanid period those of the Masjed-e ʿAlīšāh, whose massive qebla wall dominates the city’s skyline. For a long time the remains of the latter mosque were used as the citadel (arg) of Tabrīz (see ARG-E ʿALĪŠĀH); after the Revolution of 1979 they were restored to their original purpose and made into an enclosure for public prayers in the open air.
Parts of the north of Iranian Azerbaijan were inhabited by Armenians before the mass expulsions and emigrations of the First World War. Numerous churches and ruins of churches attest the density of this population, particularly in the area northwest of Lake Urmia. Some are of considerable artistic and historical interest, such as the church at Mojombār near Tabrīz which probably dates from the 4th/10th century. The church of St. Thaddeus, locally called the Qara Kelīsā, on the site of the saint’s tomb is partly 4th/10th century (the east end); it was largely rebuilt after an earthquake in 1318 and greatly extended in the 13th/19th century. There were hopes on the Iranian side at that time that the monastery of St. Thaddeus, if suitably enlarged, might become the seat of the Catholicos, but when political factors rendered these hopes vain, the nineteenth-century building work was left unfinished; even so, this is one of the most interesting Armenian churches. Equally noteworthy is the monastery of St. Stephanos on the frontier-river Aras; parts of it go back to the 3rd-4th/9th-10th centuries, but most of what remains today dates from the 10th-11th/16th-17th centuries and gives interesting evidence of mutual interactions between Christian and Islamic art in that period. (For reports of surveys of Armenian churches in Azerbaijan, see AMI, N.S. 2, 1969, pp. 8ff., 12, 1979, pp. 361ff.; on St. Thaddeus, Documenti di architettura armena 4, Milan, 1973; on St. Stefanos, ibid., 10, 1980.) Armenian influence on Iranian architecture is apparent in the gateway of the bazaar entrance to the town of Ḵoy, probably built by Armenian masons in the early 13th/19th century when the town’s walls were broadened to form “French-type” fortifications; the use of alternate layers of different-colored stone is typical of Armenian stonework.
The most impressive relics of Safavid architecture and fine art in Azerbaijan are the tile-clad mausoleum of the dynasty’s founder Shaikh Ṣafī at Ardabīl and the adjoining chamber which was built to house the royal collection of Chinese porcelain. Lesser buildings of the Safavid and likewise the Qajar period in Azerbaijan give the impression that the province made no significant cultural advance in the 11th/17th and subsequent centuries. When the central government left Tabrīz, Azerbaijan was relegated to a subordinate role and all the artistic talent of Iran was drawn to the succeeding capital cities, first Qazvīn, then Isfahan, and finally Tehran.
Trade routes across Azerbaijan have long been important. Stretches of some of them were described in ancient times by Ptolemy. Economic growth in the Middle Ages gave rise to increasing intercontinental traffic, particularly on the branch of the “silk road” crossing Azerbaijan from east to west. Although in general nothing much was done in Azerbaijan to improve the state of the roads, construction of bridges and caravanserais was essential.
The Aras was bridged at a number of places. Remains of a Qajar (19th century) bridge on the Tabrīz-Baku road and of a probably Safavid (17th century) bridge at Ḵodā-āfarīn survive in fairly good condition (see AMI 18, 1985). Also well-preserved is the remarkable walled approach ramp of another probably Safavid bridge over the Aras west of Julfa (Jolfā) on the road from Tabrīz to Yerevan (Īravān). A bridge east of Mākū bearing an Armenian inscription is still intact. The old bridge near Tabrīz and another near Āḏaršahr deserve mention, and the bridge called the Pol-e Qāflān-kūh over the Safīd-rūd near Mīāna is particularly interesting as an inscription records the date of its completion in 888/1484 (Kleiss, AMI 16, 1983, pp. 363ff.). Dating of bridges which lack an inscription is difficult, but in general it appears that the surviving old bridges or ruins of bridges date from the Qajar period (19th century) or less frequently the Safavid period (17th century).
Azerbaijan is not so rich in caravanserais as are the areas around the central Iranian desert, the Dašt-e Kavīr, and along the cross-desert routes. The main road from Yerevan through Jolfā, Marand, Tabrīz, and Mīāna to Qazvīn and Tehran was endowed with Safavid and Qajar caravanserais as well as the Jolfā and Pol-e Qāflān-kūh bridges. Remains of the tile-adorned portal of a Timurid (14th-15th century) caravanserai can be seen between Jolfā and Marand (Kleiss, AMI, N.S. 5, 1972, pl. 53.3). Most of the caravanserais are of the courtyard type with four ayvāns (arched portals), the commonest form of caravanserai in Iran. On roads which cross the frontier to Turkey through high passes, particularly the Tabrīz-Ḵoy-Van road, some caravanserais of the completely covered type remain, and there is also one close to the pass on the Tabrīz-Ahar road; they were built to give shelter from avalanches and snowstorms and therefore have no courtyard, all the rooms and stables being interconnected and vaulted.
Azerbaijan has a large number of medieval castles, mainly in mountainous areas. Many stand on pre-Islamic fort sites. They were built to guard and control lines of communication, to overlook and protect cultivated areas, and to dominate cities and towns. Certain castles, such as the Qaḷʿa-ye Doḵtar above the Pol-e Qāflān-kūh, have architectural peculiarities which suggest that they were built by the Assassins to secure the communications between their headquarters at Alamūt northeast of Qazvīn and their outposts in Syria.
Much work remains to be done on the archeology of Azerbaijan from pre-historic to recent times. Insecurity on the roads, and thereafter wars and their sequels, kept archeologists away for many decades. Between 1950 and 1978 promising excavations and wide-ranging surveys in Azerbaijan were planned and implemented in an international cooperative effort. Political developments in Iran since 1357/1979 have brought all these undertakings to a temporary halt.
Given in the text. On the pre-Islamic period, consult L. Vanden Berghe, Bibliographie analytique de l’archéologie de l’Iran ancien, Leiden, 1979, and Supplément l. 1978-1980, Leiden, 1981.
On the Islamic period, see the survey reports in Archäologische Mitteilungen aus Iran, N.S. 1, Berlin, 1968, and subsequent volumes.
See also ʿA. Kārang, Āṯār-e bāstānī-e Āḏarbāyjān I: Āṯār o abnīa-ye tārīḵī-e šahrestān-e Tabrīz, Tabrīz, 1351 Š./1972, and S. J. Torābī Ṭabāṭabāʾī, Āṯār-e bāstānī-e Āḏarbāyjān II: Āṯār o abnīa-ye tārīḵī-e šahrestānhā-ye Ardabīl, Arasbārān, Ḵalḵāl, Sarāb, Meškīnšahr, Moḡān, Tabrīz, 2535 = 1355 Š./1976.
Originally Published: December 15, 1987
Last Updated: August 18, 2011
This article is available in print.
Vol. III, Fasc. 2, pp. 215-221