CITIES

i. Geographical Introduction 

There is a long history of settlement on Persian territory, where urban life was firmly established in antiquity, and cities continued to proliferate, though, owing to fluctuations in the population, they were highly unstable. Many of those early cities have disappeared; although the ruins of a few can still be recognized, most have left no trace. Capital cities have been particularly subject to vicissitudes, each in turn rising to prominence and then being destroyed, reflecting, on one hand, shifts in political power and, on the other, the difficulties of centralization on the Persian plateau: The axis of the plateau, extending across the central and southeastern parts of the country, consists of deserts inhospitable to urban life, and cities around the periphery have tended to be oriented toward neighboring regions.

Centers and stages of urbanization. The distribution of urban centers in Persia reveals a fundamental dissymme­try. Although cities are relatively numerous in all regions where rainfall agriculture is possible, they are extremely rare elsewhere. Nevertheless, a line of oases does extend through Yazd and Kermān in the heart of the desert, constituting a particularly important link between the central plateau and cities like Isfahan and Hamadān east of the Zagros range. In the north a second chain of cities (e.g., Zanjān, Qazvīn, Tehran, Semnān, Dāmḡān, Mašhad) extends along the Alborz piedmont from Azarbaijan to Khorasan, through Herat and Qandahār south of the central mountains of Afghanistan, and as far as Ḡaznī and Kabul. A third important urban group comprises the ports and inland cities of the southwestern Zagros piedmont, where an extension of the Mesopotamian plain has always been more or less under Persian political and cultural influence. Finally, along the northern margin of the Hindu Kush there is an almost continuous series of urban centers, which in certain periods furnished the bases for a general Iranian expansion into southern Central Asia. The vitality and fortunes of these diverse centers of urban life have been extremely variable, owing to political and economic factors, combined with the spread of Iranian domination in peripheral areas.

Because of its pivotal position the northeastern perim­eter of the Zagros would seem to provide ideal conditions for the establishment of Persian capitals. Even the names of cities like Hamadān (OPers. Hamgmatāna > Gr. Agbatana, Ecbatana “meeting place”; D’yakonov, p. 277) and Isfahan (< *çpāḏāna “army camp”; cf. Ptolemy 6.4: Aspadane; Spiegel, I, p. 100) reflect a tendency toward centralization. Yet the Persian capital has never been permanently established in the region. In the 8th-7th centuries b.c.e. the Medes did establish their political center in Ecbatana, sheltered by the mountains and conveniently situated vis-à-vis Aššur, the Assyrian capital, which lay farther west, on the Tigris. On the other hand, their Achaemenid successors, masters of an empire that encompassed the fertile Mesopotamian plain and eventu­ally Asia Minor, established their capitals farther south. In the Parthian period Ecbatana was reduced to the rank of summer capital (Strabo, 11.13.1), and the main centers were Susa and Ctesiphon, in the lowlands. Isfahan served as the Persian capital under the Saljuqs in the 5th-6th/11th-12th centuries and the later Safavids (1005-1135/1597-1722), but both these episodes were brought to an end by the prevailing centrifugal tendencies of Persian urbanization. Yazd and Kermān were too far from the prosperous humid parts of the country to serve as capital, though Kermān was the seat of a local branch of the Saljuq dynasty in the 5th-6th/11th-12th centuries.

In all periods urbanization was weakest in the south and southwest. Persepolis and Pasargadae, the successive Achaemenid capitals of Fārs, were primarily ceremonial sites where people assembled for the great festival of the Mazdean new year, rather than genuine urban centers with multiple functions; they owed their status entirely to the fact that the dynasty had originated in the region. Shiraz was probably no more than a village before the Islamic period. Certainly it experienced its greatest development after it had supplanted ancient Eṣṭaḵr, but it has hardly ever been more than the regional capital of Fārs; its role as center of the short-lived Zand dynasty in the second half of the 12th/18th century was entirely owing to ephemeral historical circumstances (Clarke, 1963). Access from the Gulf to the Persian plateau was provided by a series of ports (Sīrāf, Hormuz, and Bandar-­e ʿAbbās(ī) and Būšehr), but none took permanent root before the modern period. The modern revival and development of the cities of Ḵūzestān, for example, Ābādān and Ahvāz, has essentially reflected 14th/20th-century economic realities: the exploitation of oil, which has provided a solid basis for regional expansion, and the reorientation of foreign trade toward the Persian Gulf after construction of the Trans-Iranian railway.

In fact, for about a millennium it has generally been the northern perimeters of the country, or at least the territory bordering the mountains on the east and south, that have attracted political power and thus urban development. The establishment of the Qajar capital at Tehran in the late 12th/18th century was only the latest manifestation of what can be considered a permanent tendency in Persian urban life. Many factors have contributed to the establishment of successive dynastic capitals in these northern regions, especially along the great invasion route following the base of the Alborz as far west as Azarbaijan, but the Turkish and Mongol origins of most ruling dynasties in Persia have certainly been the main determinant. Most of those rulers sought to be near the nomad tribes that furnished the bases of their power and at the same time to avoid the excessively hot climate of the south (Planhol, 1974). The development of Tabrīz, for example, can be explained by the fact that between the 7th/13th and 10th/16th centuries, despite many vicissitudes, it was succes­sively the capital of the Il-Khanids, the Qara Qoyunlū, the Āq Qoyunlū, and the Safavids, all either descended or drawing their support from northwestern Turkman tribes. During that period it was only briefly abandoned for other cities in the same region: Marāḡa, near the fertile pastures of the Sahand, which were no doubt the attrac­tion for the Mongol khan Hūlāgū (Hülegü; 654-63/1256-­65); Ardabīl, cradle of the Safavids; and Solṭānīya, founded by Arḡūn Khan (683-90/1284-91) on a grassy plateau suitable for summering flocks.

Other factors also reinforced the gravitational pull of the north. For example, the Safavid capital was withdrawn to Qazvīn during the great wars with the Ottomans in the 10th/16th century, and in the 12th/18th century Nāder Shah (1148-60/1736-47) located the center of his empire at Mašhad, reflecting his orientation toward Afghanistan and India in the east. Ardabīl and other northern cities originally grew up as frontier posts in the “marches” facing the mountains and the Caspian forest; Qazvīn, for example, was long a Muslim base against the Zoroastrian tribes of Daylam and, through almost the entire medieval period, against the heterodox Assassins, holed up in the mountains. On the other hand, in the Caspian lowlands themselves an urban network had not developed before the modern period (Kopp, 1978), and Rašt can hardly have existed before the end of the 7th/13th century. For a long time economic factors also favored the northern cities. In medieval and modern times most commercial traffic across Persia has followed the route north of the desert through the Alborz foothills. Furthermore, the first modern links between Persia and Europe were established via the Black Sea and particularly the Caspian; until the beginning of this century Tabrīz remained the princi­pal emporium, the departure point for caravans to Trabzon and Anatolia, and the terminus of the first Persian railway, which after World War I was linked to the Russian rail network.

In Afghanistan, on the other hand, cities have generally been concentrated south of the great curve of the central mountains. The early cities located on the northern slopes of the Hindu Kush were destroyed during the medieval Turkish and Mongol invasions, to which they were the most exposed and from which they never fully recovered; Bactra, for example, is only a very small hamlet today. Since the Middle Ages the capitals of Afghanistan have always been located in the south and oriented primarily toward the east; for example, in the 4th-5th/10th-11th centuries the conquest of the lowlands of northwestern India was launched from Ḡaznī. In keeping with this tendency, the rise of the Afghan tribes since the 12th/18th century and the dominant political role of the Pashtun tribes in the modern period have provided the geopolitical base for cities like Qandahār and Kabul.

Cultural strata in urban physiognomy and organiza­tion. A thoroughly original city type developed on Persian territory. It was reshaped to a considerable degree by new religious and social conceptions after the Islamic conquest but has preserved many earlier features that distinguish it unequivocally from the Islamic cities of the Fertile Crescent or India, with which, however, it shares certain elements. In order to define the Persian city and its regional variants, it is necessary first to identify these different components and prevailing currents of influ­ence.

It is certain that initially urban life on the Persian plateau was modeled on that of neighboring Mesopotamia, especially after the Assyrians and Babylonians had established fortified outposts in central Persia, in the vicinity of the future Ecbatana (D’yakonov, esp. pp. 267, 276-77). Certain material components of modern Persian cities can be traced to this early link with the Mesopotamian cultural area; for example, such household arrangements as the wind tower (bādgīr) and the sardāb (underground chamber) were already known in ancient Assyria and Babylonia (Badawy). Very soon, however, urban devel­opment in Persia took a more original course.

Foremost among indigenous elements were the sites; first settled villages and eventually cities were closely linked almost everywhere in Persia with sources of water. Cities could be on watercourses, and one entire group, very likely the oldest, comprised cities located at the points where more important streams flowed from the mountains onto the central plateau or the piedmont. Isfahan (Spooner, 1974), Ahvāz, and similar settlements belonged to this group, as did several foundations of the late Sasanian period, like Bam. By far the greatest number of cities, however, were watered by qanat@s/kārīzes (un­derground water channels). They were generally located on gradual slopes some distance from the mountains where the qanāts originated, so that the shallower incline of the plain would, through simple gravity, eventually distribute the water over the ground. The open water channel (jūy) was a standard feature of the traditional Persian street. The technique of building qanāts was common everywhere in Persia from the Achaemenid period and thus became almost the exclusive instrument for the development of cities. Owing to their extraordi­nary mastery of this type of construction, combined with ignorance of the technique of collecting rainwater in roof cisterns, Persians generally did not seek other types of urban sites in antiquity. They failed, for example, to exploit the urban possibilities of fortified high places (Planhol, 1983), even though the seeds had been planted at sites like proto-Median Sialk (10th-8th centuries b.c.e.) and Masjed-e Solaymān (seat of the first Achaemenids; Planhol, 1974, p. 150). This tradition did continue for a time in the citadels of Parthian cities and in the location of Zoroastrian cult places on cliff tops (Planhol, 1983, p. 100), but the need for a nearby source of water very quickly became the determining factor in the foundation of new urban settlements. The city of Europus, founded by Seleucus Nicator on a spur at the westernmost extrem­ity of the Anti-Alborz above Rhages, reflected a renewed tendency to seek high places after the Greek conquest, but it was on the neighboring plain that the city of Ray developed, with only the eastern end of its surrounding wall constructed on a hill (Bobek, 1958, p. 15). The evolution of this Persian city type was not sudden, probably extending over more than a millennium, parallel to the geographic spread of the technology of qanāt building. It culminated in the great period of prosperity and growth under the Sasanians, when urban colonization of the interior plateau was completed as far as the most arid zones of eastern Persia; the new foundations were exclu­sively cities of the plain, linked to water sources and most often fed, like Kermān (English, p. 22), by very deep qanāts. The pattern of the Persian city appears to have been definitively established by this period.

From this pattern arose a number of other distinctive elements, foremost among them the division of land into elongated rectangular parcels adapted to take maximum advantage of the flow of water on an irrigated slope. The lower portion of the plot was typically occupied by a garden laid out geometrically amid a system of intersect­ing canals with branches of unequal lengths flanking a pool, an arrangement that had aroused the admiration of the Greeks as early as the conquest of Alexander (Quintus Curtius 7.2; Sackville-West; Wilber). On the upper part of the rectangular plot the house was situated transversally and usually open to the garden through a veranda sup­ported on columns or pillars; sometimes this central unit was flanked by projecting lateral wings. This type of house was quite different from the traditional type with an interior courtyard and multiple patios surrounded by rooms, which had developed in the earliest Mesopotamian and Near Eastern urban centers. The Persian house developed completely independently of the rural dwell­ings characteristic of the Neolithic and the Bronze Age and at a fairly late date. It was probably derived from the basic module of the qaḷʿa type of walled dwelling, a cell incorporating a section of the fortification wall and opening toward the interior, which had evolved in southern Central Asia and Persia in response to the appearance of large-scale warrior nomadism during the 1st millennium b.c.e. (Rosenfeld). It thus represented an adaptation to the shape of the parcel on the irrigated slope (Planhol, 1974, pp. 151-52). Placing the house at right angles to the longer axis provided for necessary natural ventilation during the summer, when the oases are cooled by moun­tain breezes. It thus incorporated more ancient, originally Mesopotamian means of combating the heat (see above).

On the other hand, this type of house was poorly adapted to the type of secluded family life characteris­tic of ancient Persia and even more rigorously enforced in Islamic civilization. The demand for seclusion could be satisfied by constructing a second building (andarūn) behind the first (bīrūn), separated from it by a small strip of the garden; this rear structure was reserved for domestic intimacy, whereas the reception rooms were to be found in the main building. This solution was adopted for the more comfortable residences, but in humbler homes there was a tendency to develop a closed interior courtyard approximating that of the ancient Near Eastern house, though the shape of the plot ensured its distinctive Persian character. Among religious minorities like the Zoroastri­ans concern for defense and security sometimes led to analogous closed arrangements (Boyce). Nevertheless, the urban Persian house remained fundamentally an open structure, not sharply distinguished from the garden and watercourses. It is easy to see why Persian cities, with their flowing streams and greenery, attracted the Turks and Mongols, appealing to their taste for simple rustic surroundings. Such cities corresponded more closely to the inclinations of these rulers than did the enclosed, densely populated acropolises of Hellenized Asia Minor. They help to explain the resilience of Persian civilization and its assimilative power over the Turks, in contrast to the cultural break that followed the Turkish invasions of Asia Minor (Planhol, 1974, pp. 157, 159-60).

As for the urban plan, it was less original. Ancient forms of urban organization had been conceived according to cosmological schemes; although the typology of these schemes has yet to be established (Aubin, p. 72), it is clear that Persia shared this orientation with neighboring cul­tural areas, the influence of which can sometimes be traced through a complex chronological sequence. The oldest type consisted of circular plans, which were com­mon to both the ancient Persians and the Hittites; accord­ing to the somewhat legendary account of Herodotus (1.98), an example was attested at Median Ecbatana. This type was distinct from the quadrangular plans that domi­nated both ancient Mesopotamian and Hellenistic urban centers; after the Alexandrian conquest the Greeks intro­duced this type to the Persian world. Although the Parthians returned to circular plans, the Sasanians once again favored quadrangular cities: Ardašīr I (224-40 c.e.), founder of the dynasty, built Fīrūzābād in Fārs on a circular plan, but later in the 3rd century Bīšāpūr, founded by his successor, Šāpūr I (240-70), was laid out as a quadrilateral (Planhol, 1974, pp. 152-53). The regular plan with a geometric grid has remained predominant in Khorasan, southern Central Asia, and Afghanistan into the modern period (Belenitskiĭ et al., passim). It even inspired the choice of plan for the refoundation of Qandahār as the capital of Afghanistan by Aḥmad Shah in 1160/1747 (Fischer, p. 148), perhaps partly owing also to its similarity to the layout of the Pashtun nomad camp (Masuch). In that eastern region it certainly owed nothing to Hellenistic models or to a hypothetical later contribu­tion of the Arabs but can be explained by the penetration of cultural influences and urban conceptions from India (Gaube,1977, pp. 222-24), without entirely excluding the possibility of imitation of a Chinese model (Giese, p. 59). In this instance it serves to define an “eastern province” within the Persian cultural area (Giese), in contrast to a “western province,” where the urban grid was largely supplanted in the Islamic period. In modern times, with the extensive growth of cities, grids have reappeared, oriented in varying degrees to already-existing irrigation schemes (Bonine).

Analogous currents and conflicts of influence can be recognized, or at least suspected, in the internal organization of urban space and the differentiation of quarters. In particular the case of the specialized commercial quarter (bāzār) is now well understood. It was unknown to the Persians in the time of Herodotus (1.153), when their economy was still of an archaic eastern type based on statism (Rodinson); it was certainly the Hellenistic Greeks who introduced the free market to Persia (Planhol, 1987, p. 451). Nevertheless, Greek words for commercial structures were not borrowed in Persian, as they were in Hebrew and Aramaic (Rodinson, pp. lix-lxv); the word wāzār, from which Persian bāzār was derived, has paral­lels in Armenian and Sogdian and was certainly a very early usage (Benveniste, p. 126). The persistence of the Persian term might suggest that the commercial quarter had originally evolved independently from Greek innova­tions, but unfortunately archaeology has so far provided no confirmation. Written sources of the Sasanian period are almost exclusively juridical (Pigulevskaya, pp. 251-61), but a reference in one inscription of Šāpūr to wāzārbaḏ (lit. “chief of the bāzār”) suggests possible influence of the Greek agoranome, though the precise attributes of the Persian official are unknown (Chalmeta Gendron, p. 287). It thus seems probable that the Persians took the thing without the word and that wāzār, like sūq in the Semitic world, was simply a translation of agora (Planhol, 1987, p. 451). In the Islamic period, when the congrega­tional mosque functioned as the focal point of an integrated merchants’ quarter, the bāzār reached a high level of development, at Isfahan as early as the 5th/11th century (Nāṣer-e Ḵosrow, tr. C. Schefer, Paris, 1881, p. 352; cf. Planhol, 1987, pp. 473-74). The bāzārs of central Persia were important loci for urban innovation (Wirth; Gaube and Wirth), particularly in the Safavid period. By contrast the integrated central bāzār existed only in very imperfect form in Afghan cities, where it had to be adapted to the old Indian urban plan, with its intersection of major thoroughfares at the center, and where the separation of residential and commercial activities was never completely achieved (Wirth, pp. 31-36).

Research on the residential quarters of Persian cities is considerably less advanced, and the array of questions raised on this subject over the last quarter-century (Aubin, pp. 71-73) remain almost entirely unanswered. Analysis of the only published map of the residential quarters of a large city (Isfahan) in the international scientific literature (Schafaghi) reveals a toponymy based largely on descriptive, rather than ethnic and communal, names, an index of the adaptiveness and prolonged functioning of the urban crucible. In the respective dispositions of the royal quarter and the city proper there is again a major contrast: In western and central Persia the citadel or royal quarter (arg) is walled off within the walled city itself. In the “eastern” province (Khorasan and southern Central Asia) it is a distinct entity, an interior town (šāhrestān) clearly set off from the city proper or sometimes even lying opposite the latter at some distance (rabaḏ; Giese; Gaube, 1977). This contrast remains unexplained.