Water Provision in Early Islamic Cities: A Multidisciplinary Approach to Urban Water Governance

Abstract

The need to supply water to the cities of early Islam, and share the collective resource between different urban users, created a fundamental relationship between rulers and citizens that required both top-down and bottom-up participation. Due to the limited nature of surviving textual and material sources of evidence, however, exploring the details of how this was achieved in early Islamic cities is a difficult task which requires the integration of both textual and archaeological data. This chapter considers key aspects of early Islamic cities, including their urban plan, social structure and organisation, to provide a holistic understanding of how water was supplied to, managed within and used by individuals in these urban settlements. While the roles of elites in planning, financing and constructing hydraulic infrastructure are paramount in the textual sources, several strands of evidence point to the roles played by both individual households and larger communities, such as tribal groups and military units, who must have co-operated to manage the water supply collectively. Though major infrastructure was predominantly planned and implemented by state authorities, often in cooperation with elites, the distribution of this water at the micro-scale frequently seems to have been mediated through communal action and private, often commercial, initiative.

1 Introduction

In urban settlements in arid regions, the provision and management of water has always been at the forefront of dialogues between authorities and citizens. Indeed, exercising control over water to expand agricultural production—and the associated requirement to organise labour, maintenance and water rights—has long been viewed as one of the key drivers behind early state formation and the emergence of institutions, bureaucracies and governance structures.¹ More recent studies have criticised the supposition by scholars such as Wittfogel concerning the relationship between water management and so-called ‘despotic rule’, but only recently have water historians begun to formulate alternative pathways to analyse the relationship between state formation and water management.² By the medieval period, urbanised populations had existed in water scarce regions such as Mesopotamia for more than 4000 years. In such geographic zones, therefore, the profound challenges of providing water to densely agglomerated groups of people—as well as allocating access to different users, organising the maintenance of the hydraulic infrastructure and dealing with issues such as sanitation and pollution—were age-old problems. The rise of Islam in the seventh century, however, and the forms of urban settlement that took hold in the regions that fell under the sway of the Islamic caliphates, brought together different configurations of people with new institutional structures—often in settlements located in previously unoccupied places. Furthermore, recent scientific studies suggest that fluctuations in the availability of water, as a result of changing climatic and environmental conditions, may have been factors both in the initial emergence of Islam³ and possibly also in some of the settlement patterns that followed in its wake.⁴ Accordingly, this chapter aims to combine both the available textual evidence as well as surviving material traces, to analyse how the provision and management of water was organised within the cities of early Islam—a story both of continuity from what had existed before as well as the implementation of new ideals and social norms.

As a complicated topic that involves many different scales of evidence, spread across a large geographic region, this chapter gives a broad overview of water management in early Islamic cities before exploring certain areas in more detail. What follows, therefore, initially considers the organisation of early Islamic cities—and how the forms of urbanism that emerged in the post-conquest era potentially influenced the provision and governance of water—highlighting the significant role that tribal authority may have played in regulating the water supply at a communal level. Secondly, the chapter explores various aspects of the hydraulic infrastructure in the caliphal capitals—some of the largest cities for which significant bodies of evidence are available. Finally, the limited evidence for how water was supplied to and managed within individual domestic households in several early Islamic cities is explored. A common theme running throughout what follows is evaluating who, and which groups, exercised control and responsibility over the construction, maintenance and overall management of urban water systems and especially exploring the interlocking roles of figures representing the state, commercial interests, local communities and the private interests of both wealthy landowners and individual households.

2 Early Islamic Urban Organisation

One of the most notable facets of the new cities established during the early centuries of Islam were their relatively large populations. Though precise figures are unavailable, estimates for the population of Basra suggest a figure of up to 250,000, as many as 500,000 for Baghdad during its zenith in the late eighth and ninth centuries,⁵ a maximum of 200,000 for Samarra at its peak⁶ and a similar number for Fusṭāṭ by the early tenth century.⁷ Not only would these large populations have required a reliable supply of drinking water but also water for a wide range of secondary activities, from the mundane—laundry, bathing, food preparation and sanitation—to more specialised applications—milling, brick and pottery production and, by the eighth century, the making of paper. Furthermore, not only was water a fundamental requirement for wuḍūʾ—ablutions performed before prayer—but water was also used extensively throughout Islamic cities for purposes that were not strictly practical—such as impressive fountains, pools and the irrigation of gardens.⁸ Given both the basic necessity, as well as the economic, social and religious significance of water to urban settlement, therefore, it is difficult to imagine that the supply and distribution of water was not, at least to some extent, planned and overseen within early Islamic cities by central authorities.

The view that, at least in the immediate aftermath of the conquests, Islamic cities lacked any central planning and developed more or less organically—to become both haphazard and unsanitary—was widespread throughout the twentieth century.⁹ Scholars such as Gustave von Grunebaum discussed the concept of a so-called ‘Islamic’ city, an urban type considered to be fundamentally different from the European premodern city.¹⁰ Their ideas were nourished by Orientalist notions of a dichotomy between East and West and defined in juxtaposition to Max Weber’s conceptualisation of the ideal–typical city.¹¹ As a result, the ‘Islamic’ city has often been characterised in terms of its deficiencies in comparison to the premodern European city—particularly a perceived absence of formal urban institutions. Although area specialists have convincingly argued otherwise,¹² this assumption of institutional weakness continues to inspire general comparative scholarship.¹³ Scholarship is only recently starting to uncover what an ‘Islamic city’ was in terms of governance, though the focus has, so far, only been on specific cities or specific urban institutions and mainly in the era after the year 1000. Through his analysis of the role of urban elites in the management of local affairs in the major cities of Iraq and Egypt, for example, Mathieu Tillier argued for the existence of a ‘civic sphere’ in early Islamic cities. For example, urban elites frequently played active roles in the political life of their city, either by exercising official positions—governor, chief of police, or judge (qāḍī)—or through the periodic and informal representation of the urban community before the governor or caliph.¹⁴ This chapter contributes to the ongoing debate on urban governance and the nature of a ‘civic sphere’ by taking water as a microcosm through which urban organisation and civic participation in public services are investigated—starting with some of the urban settlements founded during the early phases of the Islamic conquests.

The cores of the cities founded during the early conquests, such as Basra and Kufa in Iraq and Fusṭāṭ in Egypt, started out as temporary military encampments which were only subsequently formalised as permanent settlements. It is probably true, therefore, that these early settlements were not planned at the outset with much concern for sustaining the large population numbers that would soon settle at these sites.¹⁵ Though non-orthogonal street plans are a typical feature of pre-Islamic settlement in Arabia and Syria,¹⁶ these were also often defined by organisation based around tribal groupings and communal, and often large-scale, water sources.¹⁷ Pre-Islamic urban settlements such as Umm el-Jimāl, Jordan, Subayta, Palestine and Yathrib, Saudi Arabia, have been interpreted as revealing evidence for internal divisions between different tribal, or clan, groups through distinct clusters of housing arranged in separate centres within the settlement. This type of organisational structure was repeated in the armies that participated in the Islamic conquests, which were principally arranged around tribal groups with kinsmen fighting together—led and administered by tribal leaders.¹⁸ Even in siting a temporary military camp, however, an accessible and reliable water supply would have been one of the primary requirements. In Fusṭāṭ, for instance, the banks of the Nile, from where drinking water was drawn, were protected as a communal resource—rather than being granted to a specific tribal group. Furthermore, the proximity of different land grants allocated to different groups and the river was probably a significant factor in the demarcation of the city’s different districts as well as the development of east/west routes between these neighbourhoods and the Nile—along which water must have been carried.¹⁹ This kin-group based social structure, therefore, directly influenced how the so-called ‘garrison towns’, or amṣār, were planned out and their internal organisation.²⁰ Furthermore, within these individual city sectors, or districts, tribal groups may have exercised a certain degree of autonomy in the construction and management of the neighbourhood²¹—which most likely included the management of water.

3 Tribes and Water Management

The importance of tribal social structures in relation to the management of water during the early Islamic period is most widely attested in rural contexts. In Oman, for example, the pre-existing tribal makeup of society strongly influenced the development of water law, property rights and dispute mediation relating to water.²² Tribal affiliations were also paramount in the villages of the Fayyūm and, by the Ayyubid period, played an important role in landholding and taxation as well as the division of water rights—which were most commonly assigned collectively to clan groups occupying specific villages.²³ In al-Andalus too, tribal organisation was integral to the management of irrigation systems—Glick explains that ‘norms of tribal governance are embodied in both the design of allocation procedures and that of the physical layout of the system[s]’.²⁴ The settlement of the Balearic Islands in the tenth century, which has been studied in detail by Kirchner, reveals how different Arab and Berber clan groups colonised areas based on the available water resources, apparently managing the irrigation systems they created largely without the involvement of higher authorities.²⁵

Since water management in rural areas was evidently often organised in close association with the tribal makeup of society, it would be natural to assume that this also played an important role in Islamic urban settlements. One case in which this may be apparent comes from the early history of Basra. Shortly after the city’s foundation, a neighbourhood was assigned to the Asāwira—a military unit of Persian origin who aligned themselves with the conquerors and converted to Islam. In various accounts retold by the ninth-century historian al-Balādhurī, it appears that the Asāwira may have had to excavate a new canal in order to bring water to the quarter of the city to which they had been assigned.²⁶ Districts, or quarters, associated with specific tribes or ethnic groups still existed at Basra when Ibn Baṭṭūṭa visited the city in the mid-fourteenth century—with quarters of the Banū Hudhayl and Banū Ḥarām (Arab tribes) and ‘the Persians’, each with their own ‘chief’.²⁷ While the Asāwira were not technically a tribe, they were assigned their own neighbourhood in accordance with the tribal division of Islamic society at the time. The brief anecdote about the canal that supplied their district advances the possibility that, at least in the period shortly after the foundation of Basra, canal digging—and by extension the construction, administration and maintenance of the water management system more generally—was organised at the level of these distinct city neighbourhoods.

4 Cooperation Between the State and Local Actors

While organisation and management within tribal units were probably important, especially at the micro-scale, state authorities²⁸ certainly also exerted significant influence and were often responsible for the establishment of major pieces of hydraulic infrastructure. Al-Balādhurī’s account of the founding of Ramla, Israel, by the Umayyad governor Sulaymān b. ʿAbd al-Malik in the early eighth century provides one such example. Here, a cistern was reportedly among one of the first things to be constructed in the city and, after a mosque and a palace had been built, the governor also established canals and wells to provide water for citizens moving to the city.²⁹ Although al-Balādhurī implies these new residents were left to their own devices when it came to the construction of their homes, the archaeology reveals that housing was carefully planned on a grid following the same alignment as the central mosque.³⁰ We may imagine here the involvement of surveyors much like those described in the Geniza documents who, in twelfth-century and thirteenth-century Cairo, ensured that land was properly divided between properties, both prior to construction and in order to settle disputes.³¹ A corresponding picture can be seen after the Islamic conquests in North Africa where, though re-occupation of pre-existing Byzantine settlement was typical, new cities were also established. Kairouan, Tunisia, for example, was founded in 670 CE and closely followed the model of the amṣār—most likely with an orthogonal street plan, at least in the centre, and separate districts assigned to different tribal groups. Major hydraulic infrastructure, including an aqueduct, reservoirs and cisterns were built under Hishām b. ʿAbd al-Malik, who ruled between 724 and 743 CE.³² Similarly at ʿAnjar, an early eighth-century foundation located in present-day Lebanon, between Damascus and Beirut, Hillenbrand hypothesises the existence of an overarching scheme prior to the construction of the urban centre which probably included the city’s hydraulic infrastructure—and specifically the removal of sewage and the intensification of agriculture in the rural surroundings.³³ Such a plan may have been achieved using either a pictorial depiction—for example drawn on animal skin—or markings laid out on the ground—such as ashes arranged in dark lines—to plan the form and location of these civic features prior to construction.³⁴

Though these examples all clearly highlight the importance of the state, and often the caliph himself, in the design and implementation of hydraulic infrastructure, this may be a misleading picture. Within the rectilinear symmetry of ʿAnjar, for example, shown in Fig. 1, which presents a clear picture of order and authority imposed from above inspired by Roman antecedents, the internal organisation appears to have been to some degree compartmentalised—with discrete blocks given over for specific purposes, such as domestic housing in one area and official and religious structures in another. Possibly, in a similar way to the tribal districts of the amṣār, each sector may have been to some extent autonomous—planned and administered in isolation.³⁵ One reading of ʿAnjar, therefore, which may be applied more generally, is of a skeleton plan laid out at the outset by a high authority, that included the major municipal infrastructure. Subsequently, within this overarching schema, lower-status actors and groups were given leeway and autonomy to exercise control within specific districts. Such a model accords with what Hugh Kennedy describes as the ‘minimalism’ of the early Islamic state which was generally happy to let local actors and communities regulate themselves so long as this presented no threat to wider affairs.³⁶

Fig. 1

The walled town of ʿAnjar, Lebanon, with major water management features marked. Redrawn by Peter J. Brown after Finster (2003)

A comparable approach—in which top-down planning and investment provided the impetus for management and extensification by local actors—may be apparent in the major irrigation schemes enacted in rural areas. In Syria, for example, there are numerous examples of large canals constructed by the Umayyad caliphs and princes. Some evidence suggests that major hydraulic works such as those along the Euphrates, and along other major rivers, were regarded as the exclusive responsibility, and remit, of the state and were funded by the central treasury.³⁷ Similarly, many of the largest canals that existed in the hinterland of early Islamic Basra bore names which, either directly or indirectly, referenced either the caliphs themselves or the governors of Basra³⁸—with further textual evidence often attesting to their direct involvement in the construction of these canals.³⁹ In such cases, however, rather than the provision of water as a public service, private investment by members of the elite was often motivated by commercial interests and Islamic landholding practices—which assigned ownership to the individual who made ‘dead’ land cultivable.⁴⁰ Furthermore, as both Eger and Verkinderen have noted, there is considerable bias in the historical sources as infrastructure which was financed and constructed by members of the elite—which were usually the largest and most important canals—is significantly more likely to be described and recorded in the surviving textual sources.⁴¹ More minor canals, on the other hand, which are more likely to have been community initiatives, dug and maintained through local collaboration as offshoots from these larger more substantial canals, rarely appear in the historical sources. A rare example in which the maintenance of hydraulic infrastructure is discussed in detail comes from the Fayyūm in thirteenth-century Egypt, where a complex picture emerges in which the wealthy landowner provided significant financial investment to restore the system, but the work was guided and carried out by the local peasantry who were familiar with, and reliant upon, the irrigation infrastructure.⁴² In such cases, it is evident that a complex bilateral relationship existed between state authorities, or wealthy landowners, and the local communities that depended on these systems.

5 Water in the Caliphal Capitals

A multi-layered dialogue between ‘public’ and ‘private’⁴³ interests is also apparent in water provision in urban contexts. As the caliphal capitals, such as Baghdad and Samarra, were some of the largest and best documented early Islamic cities, these make a useful case study to explore in greater depth. Identifying the varying (de)centralised and top-down/bottom-up processes in relation to water provision in these cities also more clearly highlights the deeply nuanced nature of issues such as the divide between ‘public’ and ‘private’ spheres of life. Much of the evidence relating to the succession of caliphal capitals built during the eighth and ninth centuries emphasise the roles of caliphs and members of the elite in the planning and construction of at least the major ‘backbone’ of these cities’ hydraulic infrastructure. In Baghdad, for example, the new capital built in central Iraq in 762 CE, the geographer al-Yaʿqūbī (d. 897–898) relates how, during the initial stages of construction of al-Manṣūr’s round city, wells and canals were dug to bring water to the building site, not only to supply drinking water but also, for ‘brick-making, and moistening clay’.⁴⁴ Unfortunately, the layout of early Islamic Baghdad is not known through physical remains and, though textual sources describe the configuration of the canals that flowed through the city, details of their financing, construction and management is rarely discussed. In one case, Ibn Serapion mentions that a canal was dug by the caliph al-Muʿtaḍid, who ruled between 892 and 902 CE, suggesting continued elite oversight and investment in the hydraulic system. It is notable, however, that his description of the canals often relates their route to the locations of various palaces suggesting that providing water for these elite residences was often a primary goal.⁴⁵ Similarly, at Hārūn al-Rashīd’s capital of Raqqa, Syria, significant hydraulic infrastructure was constructed, seemingly both as a ‘public good’ as well as for the caliph’s own benefit—with a major canal excavated to provide water for the populace alongside a separate canal that brought water exclusively to a group of palaces.⁴⁶ This can also be seen at Samarra, where the caliph al-Mutawakkil (847–861 CE) embarked on the construction of a new city district alongside a large palatial complex to the north of the existing developments built under his father al-Muʿtaṣim (833–842 CE). One of the key elements to this new development was a canal—the nahr al-Jaʿfarī—that would bring water, firstly, to his new palaces and, secondly, throughout the new district he had planned.⁴⁷ Construction began in 859/860 CE and although, ultimately, the canal was never completed, work was ongoing for approximately two years at a cost of perhaps more than a million dinars.⁴⁸

Though elite, and especially caliphal, involvement in these cities’ water projects, as funders and planners, often dominates the surviving textual accounts of these projects, non-elite involvement in the construction of hydraulic infrastructure is also apparent. In the construction of Samarra, for example, al-Yaʿqūbī describes how al-Muʿtaṣim recruited specialists who were skilled in ‘channeling and measuring (the flow of) water, tapping water, and finding underground water’.⁴⁹ Ibn Serapion, meanwhile, mentions the presence of a ‘quarter of the canal-diggers’ in Baghdad⁵⁰ while canal diggers, under various names, crop up in a variety of eighth-century to eleventh-century sources from Iraq.⁵¹ Canal construction, therefore, appears to have been a profession carried out by specialists though their skills were perhaps transferable. When Ibn Ṭūlūn decided to build an aqueduct bringing water from a pond east of Fusṭāṭ to the southern outskirts of his new development at al-Qaṭā’iʿi, in what is today southern Cairo, he recruited an architect to plan out and oversee construction.⁵² That this same architect reportedly also designed the mosque of Ibn Ṭūlūn a few years later, however, demonstrates that ‘architects’ were involved in the design and construction of all types of structure and hydraulic infrastructure was not the sole domain of specialists in hydraulic engineering.

Presumably, a great deal of unskilled labour would also have been required to enact these kinds of plans and we may speculate about whether labourers were employed as full-time canal diggers or whether corvée or enslaved labour was more common. By the time an eleventh-century treatise on canal construction was composed, full-time professionals seem to have been the norm in Iraq though the extent to which the diggers, and the workers who carried away the excavated soil to form the adjacent levees, worked on canal construction full-time is difficult to gauge.⁵³ Perhaps atypically, in the construction of the nahr al-Jaʿfarī at Samarra, 12,000 men reportedly toiled under the supervision of Dulayl b. Yaʿqūb al-Naṣrānī, a secretary to one of the commanders of the Turkish troops stationed at Samarra, to complete the canal.⁵⁴ The involvement of the military leadership in the construction of the nahr al-Jaʿfarī, alongside the high numbers who worked on the canal, advances the possibility that, at least at Samarra, the military may have provided the labour to build the hydraulic infrastructure. The figure of 12,000 given for the workforce agrees quite closely with Hugh Kennedy’s estimate of 15,000 for the maximum number of Turkish troops stationed at Samarra at any one time⁵⁵ and matches figures given for other military forces in the city shortly afterwards.⁵⁶

Certainly, the large numbers of military personnel garrisoned at Baghdad and Samarra would have required a reliable and well-managed water supply. At Baghdad, the army was housed in several different areas and, though these cantonments were not divided along tribal lines, each was assigned to a specific military unit—reflecting a conscious Abbasid policy to diminish the social order, and military connections, that had existed under the Umayyads.⁵⁷ At Samarra too, large parts of the city were given over to cantonments granted to specific military commanders, somewhat divided by the geographic origins of their troops.⁵⁸ While at Baghdad, the cantonments are only known through textual descriptions, at Samarra their locations and much of the internal layout can be reconstructed in detail based on early aerial photographs of the site. These cantonment districts account for about 40% of the city and likely housed an even larger percentage of the population due to their density.⁵⁹ These districts are distinctive for their regular orthogonal layouts with grid-pattern streets and individual units or houses arranged in rectangular blocks.⁶⁰ One of the few details about how water was managed within these cantonments comes from al-Yaʿqūbī who states that, once al-Muʿtaṣim had allocated different areas for the development of the cantonments, he built public bathhouses as well as ‘[establishing] a small market in which were a number of shops for grocers, butchers, and other essential tradesmen’.⁶¹ Though bathing facilities appear to have been, to some extent, a public utility, foodstuffs were presumably supplied to individual cantonment blocks where water would have been required for food preparation and cooking.

Fig. 2

The Abbasid city of Samarra on the banks of the Tigris. The major cantonments are marked (dark grey) and labelled. Created by Peter J. Brown after Kennet (2001) and Northedge (2007) using data from the Copernicus Digital Elevation Model (DEM) and Natural Earth

The boundaries of several of the cantonments in Samarra, such as al-Karkh, abut canals from which water could have been drawn. Presumably, as in the initial division of the tribal allotments at Fusṭāṭ, discussed above, proximity to the river would have been an important consideration in choosing the locations of the cantonments. While some cantonments are located almost directly on the banks of the Tigris (al-Dūr, al-Maṭīra, Balkuwārā and Site/Area G) others, such as the eastern part of al-Karkh, lie up to 3 km from the river. In this case, however, the area of al-Karkh furthest away from the Tigris was a later addition, possibly dating to the time of al-Mutawakkil, so its distance from the river is likely explained by the fact that more convenient sites were not available as all potential riverside sites would have already been developed by this period.⁶² As this section of al-Karkh also abuts the pre-Islamic Nahrawan canal, another possibility is that this feature provided the water supply. The extent to which pieces of hydraulic infrastructure, such as the Nahrawan canal, were maintained and managed during the Abbasid occupation at Samarra is unclear. Observations made of the upcast from some of the Sasanian-era canals, which continued to function during the Islamic period, in some cases point towards phases of cleaning but it is not known whether these works coincided with the occupation of the Abbasid city.⁶³ In a much more clear-cut case, at the city of Sultan Kala (Merv, Turkmenistan), excavation attests to the careful maintenance that took place to keep the main water supply channel flowing for about 470 years—with the lack of silt deposits encountered at the base of the canal indicating regular cleaning.⁶⁴ Though we cannot say exactly who carried out the work at Sultan Kala, it seems likely that similar maintenance works would have taken place at Samarra.

6 Residential Water Supply and Commercial Involvement

At Samarra, the importance of the water provided by the Tigris is borne out by the textual evidence. Al-Yaʿqūbī, a contemporary of the city’s short-lived occupation, maintains that drinking water for the entirety of the city’s residents was transported from the Tigris in skin bags by water carriers on beasts of burden.⁶⁵ The reason for this was that wells had to be dug to a great depth and the limited water they provided was both salty and unpalatable—though such water may still have been widely used for tasks such as cleaning, washing clothes, bathing and watering animals. While some drinking water may have been drawn from the canals and qanats that ran through some parts of the city, al-Yaʿqūbī’s account suggests that the river itself was the primary source. As much of the parts of Samarra immediately abutting the Tigris have been lost to erosion, it is impossible to pinpoint the riverside spots where the carriers obtained their water as can be seen occurring in a much more recent photograph at Mosul in Fig. 3. In transporting the water of the Tigris throughout the city, the water carriers of Samarra provided an essential service which is similarly visible in many other medieval Islamic cities—and, indeed, water carriers have been a ubiquitous feature of Middle Eastern cities up until the twentieth century. The existence of a ‘street of the water-carriers’ in early Islamic Baghdad demonstrates that, here too, many people were involved in the provision of water.⁶⁶ As in Samarra, this was presumably a city-wide service which provided drinking water to all the city’s residents. This was also the case in Fusṭāṭ where a more detailed understanding of how the water carriers may have operated is possible. Within the houses that have been excavated at Fusṭāṭ, for example, Harrison argues that niches located in entranceways, with easy access from the street, contained water jars—allowing water carriers to easily refill them to top up the household’s water supply⁶⁷—something which is also described in the Geniza documents.⁶⁸ That this kind of arrangement was practised in other urban settings may be demonstrated by the remains of one of the houses at ʿAnjar (House IV) in which a small private cistern was found adjacent to the entranceway.⁶⁹ At least in Fusṭāṭ, it seems that the water carriers were not public employees but commercial service providers who were paid directly by households and other users.⁷⁰ These water carriers were not public employees but commercial service providers who were presumably paid directly by households and other users. Exactly how this service was arranged and how it operated in practice is a matter for speculation. Perhaps water carriers were engaged to deliver water to a household at regular intervals, or on specific days, presumably with the possibility to order extra deliveries if more water was required—for example because of the presence of guests or festivities and special occasions.

Fig. 3

Water carriers collecting water from the Tigris at Mosul. Date unknown. Underwood & Underwood. Public Domain

The few excavations that targeted domestic housing at Samarra were not as detailed and, in many cases, not as extensive, as those from Fusṭāṭ though they do reveal some aspects of how water was supplied to and used within domestic structures in the city. Where external entranceways were recognisable, niches that could have been used for water storage cannot be identified⁷¹—though the practicality of having spaces and facilities close to the entranceway where water carriers could easily deliver drinking water for the household’s use makes it probable that this was a common arrangement. A typical feature of the domestic architecture at Samarra is the central rectilinear courtyards with rooms arranged around the four sides. In many cases these courtyards appear to have been foci for activities involving water as many contained drains, cisterns and basins⁷²—presumably relating to everyday tasks such as cleaning, laundry and food preparation. A number of the houses were also equipped with private latrines and bathing facilities—with these rooms distinguishable by their bitumen flooring and wall-coatings required to water-proof the mud-brick. Of the houses which contained private baths, none of these were heated by hypocausts unlike the city’s public baths indicating a key difference between the two.⁷³ Importantly, there is no evidence that any of the domestic structures were connected to anything resembling a city-wide water supply. While systems of pipes existed within individual houses, these seem to have been intended exclusively to allow water to be drained out onto the adjoining street outside.⁷⁴ At least in the small sample of houses that have been excavated at Samarra, therefore, water seems to have been supplied and managed within the unit of individual structures rather than at a more communal level.

The Geniza documents furnish us with analogous descriptions of the situation in Fusṭāṭ. Here, wells were commonly located in the central courtyards around which houses were arranged—though again this water was not used for drinking as water from the Nile was preferred. In at least one case, residents living in the upper storeys of a building had to first obtain permission from the owners of the well below to use its water.⁷⁵ Some of the pits detected across the domestic areas that have been excavated at Fusṭāṭ may be the remains of such wells.⁷⁶ Apparently in contrast to the situation in Samarra, however, Goitein argues that the sanitation system in Fusṭāṭ must have been managed by a government apparatus to oversee the installation, cleaning and maintenance of underground pipes.⁷⁷ Documentary evidence attests to payments made for the cleaning of ‘qanāts’ which Goitein suggests represent payments to a government authority that probably employed a private contractor to carry out this task.⁷⁸ It seems likely such a system might have existed in Samarra in relation to the major conduits that ran through the city—including canals and qanats. The apparently self-sufficient nature of the water systems revealed in the excavations of domestic areas, however, suggests services such as maintenance and sanitation were probably carried out at the level of the household—either through a service organised at the city or neighbourhood level or by private service providers.

Echoing the situation that appears to have been established in cities such as Samarra and Fusṭāṭ, many of the cities of the Levant that predated the Islamic conquests saw a gradual shift in the organisation of water during the Islamic era. At Caesarea, for example, the water supply seems to have experienced a gradual decline after the Islamic conquests. The public water system of Byzantine times—which included large aqueducts—may have already been falling into disrepair by the late sixth century⁷⁹ and seems to have been non-functional by the seventh century when private cisterns within courtyards became common. By the late eighth century, a new urban layout, based around grid-patterned streets had been established in which almost all houses had individual wells and cesspits with water supplied and drained via channels running alongside the streets.⁸⁰ This pattern seems to have been repeated at Jerash where the hydraulic system of the classical city was severely damaged by an earthquake in 749 CE. As a result, the aqueducts, cisterns, reservoirs and piping, which provided water for numerous fountains, bathhouses and industrial activities, fell out of use during the Islamic period. The system established after the earthquake saw no return to those of Roman and Byzantine times and instead water supply seems to have been organised at the level of individual households—with each house equipped with its own cistern located in an open courtyard.⁸¹ At Tiberias too, the earthquake of 749 led to significant rebuilding and the Islamic-era domestic architecture appears to have relied upon private cisterns while the city’s Roman-era aqueduct ceased to be maintained.⁸² The destruction caused by the earthquake, therefore, provided an impetus to move from a system that must have required significant management from public authorities to one that apparently placed greater emphasis on private responsibility—mirroring arrangements found in cities such as Samarra and Fusṭāṭ that were Islamic-era foundations.

7 Conclusions

In planning, constructing and governing the monumental cities of the early Islamic period, the textual sources emphasise the paramount role played by state authorities in the construction of hydraulic infrastructure. Archaeological evidence also provides numerous examples of these major water management initiatives which point towards the involvement of the state in the implementation and organisation of these systems. Some of these were incredibly expensive schemes requiring significant marshalling of labour and financial resources. There are important nuances to consider, however, when it comes to who organised water provision and for what purposes. When it came to constructing new hydraulic infrastructure, although the caliph often played the role of funder, planner or decision-maker, water was often used for private purposes, and the elite too were often motivated by commercial interests rather than the provision of water for the ‘public-good’. At the same time, however, a considerable body of evidence from a wide variety of early Islamic cities highlights the key role that communities and individuals played in the management and maintenance of the urban water supply.

In the immediate post-conquest period, the tribal structure of the army influenced the spatial and social organisation of the amṣār, the newly built garrison cities, and the regulation and management of the water supply was probably one of the things that was organised within the distinct neighbourhoods that formed in these settlements. A comparable decentralised system of managing the water supply within individual neighbourhoods may have also been practised in the military cantonments of later cities such as Baghdad and Samarra. This key role, apparently played at the neighbourhood level, reflects the co-operation and ad hoc organisation of services that was often prevalent in the absence of more formal intervention from state bodies—who were typically more active in the regulation of ‘public’ spaces such as markets.⁸³ An important role was also seemingly played by private service providers including not only architects and specialists in planning and constructing hydraulic infrastructure but also water carriers and, perhaps sanitation workers. These considerations place the overarching role played by caliph and state in perspective. We should also remember the evidence from cities such as Baghdad, Raqqa and Samarra, in which the water supply of palaces and elite residences appears to have been at least of equal importance to that supplied to the city’s residents. At the other end of the scale, individual households managed their own water use—perhaps through agreements with water carriers, drawing water from private wells and draining water away through drains and pipes within their own homes.

Some of the conclusions reached through this analysis of the textual and material evidence echo those made by von Grunebaum in his 1955 essay The Structure of the Muslim Town. Von Grunebaum also viewed tribal social structures and private households as important units within Islamic cities⁸⁴ though his arguments subsequently attracted significant criticism for their dogmatic repetition of orientalist tropes.⁸⁵ In the case of water management, his conception of medieval Islamic cities certainly does not match the picture assembled by the various strands of evidence considered here. He states, for example, that: ‘residents of the Muslim town did not develop their own administrative machinery’ and that ‘the administrative framework … was imposed by the state’.⁸⁶ The evidence presented here for how water was organised within various early Islamic cities, however, suggests that while major infrastructure was, indeed, planned, financed and constructed largely through state intervention, when considered as a cohesive whole, urban water systems were a much more complex web of rights, relationships and obligations between the state, distinct neighbourhoods, commercial service providers and private citizens.