Ottoman Algeria

At the beginning of the sixteenth century, the central Maghrib – the broad swathe of North Africa between the domains of the Wattasid dynasty of Fez to the west and the Hafsids of Tunis to the east – was caught in a vast geopolitical crisis. The collapse of the late medieval Muslim kingdoms of Andalus in the face of the Spanish Catholic reconquista had culminated, in January 1492, with the fall of Granada. The Spanish advance across the Mediterranean, into fortified enclaves on the coasts of North Africa, accelerated the fragmentation of the dynastic states that had ruled the Maghrib in succession to the great unified medieval empires that had passed away in the thirteenth century. The central Maghrib that would become Algeria was not yet conceived of by any of its inhabitants as a single territory, much less one ruled from the port city of al-Jaza’ir Beni Mezghenna (Alger to the French, Algiers to the English), ‘the islands of the Beni Mezghenna’, an ancient but modest harbour built against the steep relief of the hillsides facing a small group of islets off the coast, an unsheltered anchorage at the western end of a broad bay. Political sovereignty over the plains and mountains of the central Maghrib and their inhabitants was contested between the Zayyanid dynasty, ruling from the princely trading and textile-manufacturing city of Tlemcen, inland in the west of the country, the Zayyanids’ rivals in Fez further west, and the opposing force of the Hafsids of Tunis in the east, who ruled over the learned port city of Bejaïa at the mouth of the Soummam river, 200 kilometres east of Algiers. Continuous regional warfare between them was now combined with the destabilising local effects of a Mediterranean superpower struggle: the Ottoman Empire, in the ascendant since the capture of Constantinople (Istanbul) in 1453, contended with the Habsburgs, embarking on their own ‘golden age’ with the establishment and pillaging of a New World empire in the Americas. After the fall of Granada, the Spanish ‘crusade’ encroached on the central Maghribi coastline. Mers el Kebir, the great natural harbour just west of Oran, was taken in 1505, Tenes in 1508, Oran itself and Bejaïa in 1509. Two years later, Dellys, Cherchell and Mostaghanem were obliged to pay tribute to the Spaniards. All the principal coastal towns of the central Maghrib, east as far as the Hafsid domains, had fallen under Spanish control. In 1510, the notables of Algiers too sued for peace, and were obliged to cede possession of the offshore islands commanding their harbour, on which the Spanish commander, Don Pedro Navarro, built a fortress, the peñon, with a garrison two hundred strong.²

It was in this context that the notables, the chief citizens, of Algiers first called on outside aid, requesting help from an Ottoman adventurer recently arrived in the region – Aruj Bey, who along with his three brothers became better known in history and legend under the surname ‘Barbarossa’, meaning ‘red beard’. Aruj, a Muslim soldier and seaman from the Aegean island of Mytilene, had his first successes as a corsair (privateer) operating from the island of Djerba off the southern coast of Tunisia. Private naval entrepreneurs acting for their own profit but within the context of Ottoman operations against the Spanish were already a feature of warfare in the region, as they were in the Atlantic and Caribbean, where English privateers harassed the Spanish fleet. In the 1490s, the Turkish privateer Kemal Reis, later appointed to command an Ottoman squadron charged with suppressing piracy in the Aegean, had led a corsair fleet against Spanish shipping off the Maghribi coasts.³ The careers of the elder Barbarossa brothers, Aruj and Khayr al-Din, were especially successful examples of this wider trend. Having created his own fleet, Aruj was called upon to reinforce the attempt to retake Bejaïa from the Spanish. He tried, and failed, to do so twice, from his base at Jijel, but from there, as a leading anti-Spanish war captain, he began to carve out an independent political power of his own. In 1516, he was solicited by the population of Algiers to come to their aid.

But however important the regional context, this was no straightforwardly ‘civilisational’ war, with clear and simple battle lines between Muslims and Christians. Local struggles were more complex. Aruj, installed with his soldiers in Algiers, was unable to destroy the peñon, but threatened the position of the ruling faction in the city and its leader, Salim al-Tumi, who sought Spanish assistance to remove the Ottomans. Aruj suppressed this budding ‘revolt’, allegedly murdering Salim in his bath; Salim’s son sought refuge with the Spanish and hoped to avenge his father with their aid.⁴ A year later, when an anti-Spanish faction in Tlemcen invited Aruj to extend his rule and evict Spanish influence from the west, a rival group around the reigning Zayyanid sultan, Abu Hammu, opposed annexation from Algiers. With Spanish soldiers, Abu Hammu’s supporters laid siege to Aruj in Tlemcen and massacred troops commanded by his brother Ishaq at the nearby stronghold of Qala‘at Beni Rashid. Aruj resorted to extreme repression against suspected opponents in Tlemcen, murdering some seventy Zayyanid princes, according to some sources (and the local historical tradition of the city), before being killed himself while fleeing from Tlemcen in 1518. In Tlemcen as elsewhere, different aspirants to sovereignty ‘disputed power among themselves, some seeking support from the Spanish, others from the Ottomans, and changing allies as circumstances dictated’.⁵

Out of this complex local power struggle, Aruj and his brothers had succeeded by 1518 in establishing the makings of a military state, supported by factions of the local urban elites opposed to the Spaniards and backed by their small force of ships, Turkish musketeers and artillery. After the death of Aruj, his brother Khayr al-Din became ‘sultan of Algiers’. In August 1519, his combined army of local, Andalusi and Turkish troops repelled an attack by some five thousand Spanish soldiers and later that same year he offered the sovereignty of his precarious kingdom to Selim I, conqueror of Damascus and Cairo. In 1529 Khayr al-Din succeeded in ousting the Spaniards from the peñon and by 1533, when he became kapudan pasha, Grand Admiral of the Ottoman fleet of Sultan Suleyman I ‘Qanuni’ (‘the Lawgiver’, known in Europe as ‘the Magnificent’), a new state had come into being in the central Maghrib, centred on a new capital city: the Ottoman Regency of Algiers. In 1541, when storms and tenacious resistance defeated the assault on Algiers led by Charles V, the Holy Roman Emperor, with 516 ships and 22,000 men, this new state no longer seemed precarious; it began to be called invincible.

The epic, often romanticised, tale of the Barbarossas and the founding of the Regency provides in succinct summary the material for many of the misperceptions that continue to surround the three centuries’ history of Ottoman Algeria. Interpretations of the Ottoman period have tended to fall into one of two broad camps, both very much defined by anachronistic suppositions about ‘national’ development. For some, the Regency was the foundation of an Algerian national state, as yet incompletely integrated, but nonetheless establishing durable frontiers with its neighbours, exercising sovereign foreign relations from a permanently established capital and illustrating a ‘glorious’ history of warfare and resistance against Christian imperialist aggression.⁶ For others, and in contrast to the more positive models of Egypt or Tunisia, the Turkish ‘occupation’ of Algeria failed to see the emergence of an indigenised ruling elite in command of an emergent national entity. In this latter view, the ruling minority remained a foreign military clique, dependent on the revenues of the ‘piracy’ by which Aruj had begun his rise to fame and which his successors would pursue for three hundred years, and detached from local society, which it ruled as a ‘colony of exploitation’,⁷ using ruthless taxation and brute force to suppress the country’s capacities for development and oppress its populace to such an extent that the Ottomans are said to have ‘developed an incapacity to resist […] and objectively prepared the ground for French colonisation’.⁸

It is easy enough to see that both of these positions were defined by twentieth-century preoccupations. Algerian publicists addressing the international community during the later stages of the war of independence relied on arguments about the internally functional and internationally recognised legal status of the Regency to establish precedence for their own juridical claim to represent a legally sovereign state. The French government conversely insisted on its own, nineteenth-century, sovereignty by right of annexation over a territory whose people, occupied by the ‘despotic’ Turks before the ‘civilising’ French, had supposedly never achieved political autonomy and were still in need of paternal guidance towards modern development.

Understanding of the Ottoman period has also often been obscured by the very scale and colour of the Regency’s early modern history. Salim al-Tumi strangled in his bath, the massacre of the princes at Tlemcen, the exotic figures of the Barbarossas, the plots and assassinations of Algiers palace politics and above all fantastic and sexualised images of the depredations of corsair ‘piracy’, the enslavement of European men and women and graphic imaginings of the violence and licentiousness to which they were supposedly exposed, all contributed to typecast and self-reproducing images of Algiers. It was the ‘scourge of the Christian world … terror of Europe … haven of pirates, theatre of every kind of cruelty’, in the words of a French churchman in 1619; ‘this nest of wasps, this den of thieves’, as an English commentator raved in 1728.⁹ Colonial-era writers both revelled in the barbarously romantic ‘Barbary legend’ of the larger-than-life Barbarossas and their heirs, and produced durable stereotypes about Algeria as anarchic, licentious and violent. Algerian nationalists both celebrated the heroic exploits of the ghuzat al-bahr, the sea-roving ‘combattants of the faith’, and deplored the failure of the Ottoman state to indigenise itself and beat back the French invasion. Understood in terms of its own time, and in the increasing light of contemporary sources and their critical interpretation, another assessment of the Regency is possible, one that can take account of the nature of the state and of the disparate social groups and territories it sought to govern.

Living on the Land

Ottoman Algeria was an overwhelmingly rural society. Throughout this period, some 90 per cent of the population lived on the land, and the greater part of the wealth that the country generated, and whose surplus enabled the functioning of the state, was derived from agricultural production. The land itself, the productive environments it provided and the ecologies (that is, the relationships between human populations and their physical environment) produced by the interaction of human settlement with the resources of the land and climate were highly diverse. The territory that came under the nominal, and variably effective, political sovereignty of the Regency extended from the Mediterranean coast to the principal settlements of the north-central Sahara, traversing four main topographic zones.

From north to south, the first of these is a chain of coastal mountain ranges, rising to around 1,000 metres, of which the most notable are the Dahra, north of the river Cheliff, maritime Kabylia, north of the Sebaou valley, the mountains of Lesser Kabylia from Bejaïa to Collo and the Edough massif west of Annaba. These ranges close off the interior of the country from the sea and give long reaches of the Algerian coastline – the Dahra and Kabyle ‘corniches’ – the spectacular aspect of a wall of mountains with their heads in the clouds and their feet in the sea. They are, nonetheless, broken up by river valleys and low-lying coastal plains, especially the broad, flat plains extending westwards from the Cheliff valley to the south and west of Oran, the Mitidja plain south and east of Algiers, the immediate hinterland of Bejaïa and the plain of Annaba. A second zone is defined by the Tellian Atlas, a broader, deeper swathe of mountain chains extending in a shallow arc across the country: from west to east the mountains of Tlemcen, the Ouarsenis facing the coastal Dahra across the Cheliff valley, the Atlas of Blida defining the southern edge of the Mitidja, the Biban or ‘Iron Gates’, separating the region of Algiers from the upland country to the east, and the Hodna mountains, which turn south-eastward, away from the coast, dividing the rolling uplands of the Constantine region from the steppe to the south and west.

The southern edge of the Tellian Atlas marks the transition between the littoral region of ‘Mediterranean’ climate and vegetation and a third zone, the steppe of the High Plateaux. The broad, flat plateaux, narrowing to the east until they meet the barrier of the Belezma and Aurès mountains, change rapidly along their northern edge, and from north to south, between dry-farming agricultural grassland, scrubby, semi-arid pasture and arid, rocky tracts that merge into desert. South of the plateaux, a third series of mountains, the Saharan Atlas, rises to between 1,000 and 2,000 metres: from west to east, the Ksour mountains, Djebel Amour, the mountains of the Ouled Naïl and of the Zab, the Aurès and the mountains of the Nementchas. These divide the agricultural and pastoral lands of northern Algeria from the vast – and also very diverse – Sahara itself.

Each of these broadly defined zones, too, could be further subdivided. The other major topographic division important to understanding Algeria’s relief and ecology, however, cuts across these transversal bands from the southeast to the northwest and divides a generally drier, more pastoral west – with lower relief, less rainfall and fewer areas of fertile alluvial soils – from a higher, colder, more agricultural east – with soils rich in potassium and phosphates, more marked relief and higher average rainfall. The hills and plains of the west were the lands of nomadic, Arabic-speaking pastoralists, moving between the northern ‘Tell’ and the Plateaux in regular, seasonal migrations.

To the east, the mountains break up into smaller chains descending towards the Tunisian frontier, and open up into areas of rolling hills and the ‘high plains’, suited to wheat and barley growing, around the town of Setif north of the Hodna mountains. The cereal culture supported by the soils and rainfall of this eastern region, along with intensive agriculture in the well-watered mountains, has historically supported higher population levels – up to half the country’s total population – and a greater agricultural surplus than other areas.

A clearer idea of both the diversity and the salient common features of these different productive environments, the foundation of Algerian society in the Ottoman period and for long afterwards, can be gained from more detailed consideration of a few specific locales within these broad topographic zones.

Figure 1.1

Hills and plains in western Algeria, from Qalaat Beni Salama, near Tiaret. It was in a cave near this spot that the celebrated historian Ibn Khaldun is said to have begun the Muqaddima (Introduction) to his universal history, while in hiding in the 1370s.

(Author’s photograph)

The Sahara itself presents a great diversity of land forms, resources and possibilities for human settlement.¹⁰ At the furthest reaches of Ottoman claims to sovereignty lay the Gourara¹¹ (from tigurarin, ‘encampments’), a string of oasis settlements on the southern edge of the Great Western Erg (reg, a vast ensemble of high dunes), almost 750 kilometres south of the coast. In medieval times, the Gourara had been a node in the north–south and east–west roads across the northern Sahara. In the sixteenth century, it became relatively isolated, less frequented and only occasionally – at the moment of Ottoman and Moroccan expeditions in 1579 and 1581, respectively – subject to rival regional sovereignties.¹² Inhabited since antiquity, first by the autochthonous ancestors of a part of the black population later identified, as haratin (Berber, isemghan), with enslaved sub-Saharan Africans and their descendants, the oases were successively settled by ancient Berber-speaking migrants from the north, by Jewish and judaised berberophone arrivals in the second and seventh centuries AD and by nomadic Arabic speakers in the thirteenth and fourteenth centuries.¹³ The primary language of the region was tazenatit (Zenati), a Berber dialect associated with immigrants of pastoral origin who would become the principal inhabitants of the area.

Historian Lucette Valensi remarks of early modern North Africa that ‘the harsher the climate and the more irregular the resources, the more minute was the management of daily life’.¹⁴ Meticulous social regulation as well as technical ingenuity characterised life in the Sahara. Agricultural production – in small familial vegetable gardens and sometimes vast, as at Timimoun in the Gourara, date-palm groves – was made possible by the extraordinarily labour-intensive construction of underground canals (foggaras, or ifeli in Zenati), an ancient technique of tapping the water table and using natural declivities to channel the stream towards the oasis, where a system of allocation was controlled by carefully calibrated comb-like filters under the supervision of the kiyal al-ma, the water-measurer. Extended family groups, associated by lineage (lqawm), were historically recognised by their rights to cultivable plots and water allocations, by their possession of a communal, fortified habitation doubling as a secure granary and storehouse (agham) and later by their residence in distinct quarters of agglomerations of houses (ksur, singular ksar), set apart from their groves and gardens.

Figure 1.2

View over rooftops of a ksar in the Gourara, near Timimoun (T. Sahrawi).

The labour required to create and maintain irrigation systems and palm groves was largely provided by the servile haratin, although the people identified as such in the Gourara today claim a free labouring origin, if one of secondary status to that of their ‘white’, landowning berberophone neighbours, the hrar (free men). Strictly hierarchical social differentiation, on the basis of the successive migrations making up the population, was marked in descent or ethnicity, occupation and place of habitation, with haratin/isemghan (free labourers on others’ land) and ijemjan (slaves brought from West Africa south of the Sahara and their descendants) at the bottom of the scale and authoritative religious specialists, shurafa (families claiming descent from the prophet Muhammad) and mrabtin (descendants of ancestors locally recognised as saints), at the top.

A different ecological situation, and a different social system, existed to the north and east. In the high plains of the Constantinois (the region whose historic centre is the city of Constantine), as the eighteenth-century English scholar Thomas Shaw remarked, ‘they have a great command of water during the whole summer’¹⁵ – at least, in good years. Average rainfall in most of this region is between 400 and 700 millimetres per annum (400 mm being the minimum for cereal growing without irrigation), but as André Nouschi pointed out in his detailed history of rural life in the region, ‘levels of rainfall in Algeria can vary by up to one hundred per cent from one year to the next, such that the notion of “average” rainfall loses much of its relevance’.¹⁶ Drought or, equally disastrous, storms, heavy rain and flash floods – especially at the wrong time of year – were constant threats to the livelihood of the peasantry who constituted the great majority of the population. Hailstorms from December to March, and the dry sirocco wind from May to September, could destroy entire crops in the fields. So could locusts, as the French naturalist Jean-André Peyssonnel described during his journey with the Ottoman authorities of Constantine in June 1725:

Added to the instability of climate was the fragility of the soil’s fertility. Though naturally rich in phosphoric acid and potassium, without the addition of more fertilisers it usually needed to be rested to regenerate between crop cycles. The region’s optimum ecology therefore relied on extensive agriculture with a light plough on large, open fields – where they were available – combined with livestock-raising, with flocks or herds being moved seasonally by populations who often combined settled agriculture with a degree of mobile pastoralism. This combination of lifestyles was especially marked in another principally berberophone society of the east, the shawi-speaking people of the Aurès mountains that divide the plains of the Constantinois from the Sahara. The Aurès, a compact mass of mountains and valleys running south-west to northeast, marks a sharp ecological frontier, with the northern and central reaches of the massif supporting cereals and orchards in valleys and on the lower hillsides as well as pastoralism on the higher slopes, while the southern face abruptly presents an arid, Saharan aspect, with valley floors dominated by the date-palm.

Figure 1.3

The Aurès. The Rhoufi gorge, in the southern Oued El Abiod. Traditional houses on the cliff side above the palm grove (on the right of the picture) have been abandoned for the newer settlement by the main road higher up.

(Author’s photograph)

In the southern reaches of the Oued el-Abiod, one of the massif’s three principal valleys, the population divided the year between fruit- and date-growing in the valleys, with pasturing of flocks on the slopes, and removal of the whole community to the northern Sahara for winter pasture between October and March, leaving behind only a guard over the teklert, or communal granary and storehouse, access to which was usually controlled by the community’s elder women.

However fragile, though, in 1830 the cereal-growing east probably supported between two-fifths and a half of Algeria’s total population of between 3 and 5 million people.¹⁸ These mainly Arabic-speaking peasant producers occupied a variety of positions in a threefold land tenure system. Land belonged to the state (beylik), or could be held by communal groups (later referred to as ‘arsh land, associated with ‘arush, tribes) or private owners (milk, ‘possessed’ property).¹⁹

Much of the best land, on the plains and in permanently farmed estates near major towns, was owned by the bey, the Ottoman ‘viceroy’ of one of Algeria’s three provinces, either through purchase or through confiscation. Towards the end of the Ottoman period in the east, a stable system for farming such land, disposed of by the beylik as ‘azl (‘grant’) land, had developed. ‘Azl lands belonging directly to the beylik were frequently worked as ‘azl khammas, share-cropped land, for which workers were recruited from among the local tribes by an agent or steward (wakil). The peasants provided their labour while the state provided the other four factors of production – land, work animals (oxen, mules or horses), implements and seed – and the produce was divided accordingly, with the workers receiving one-fifth of the harvest (hence the term for sharecroppers, khammas, ‘one-fifth’). An account of the system produced shortly after the French conquest of the region observed that:

Given the risks to peasant livelihood in general, farming good state lands was by no means an undesirable option. The same source estimated that for the sharecropper, in a good year, ninety days of work on a well-favoured ‘azl produced ‘enough to provide for his family for a year, and he could dispose of the remaining 275 days for his own profit’.²⁰ ‘Azl could also be let to independent tenant farmers as ‘azl jabri, in which case families cultivated the land with their own resources and at their own risk. The beylik took a set proportion of the harvest (jabri, the ‘mandatory’ payment, also a term for the principal tax on produce on the region), such that in good years, both parties benefited, but in poor years only the state – which was due its fixed amount of produce whatever the weather – gained from the arrangement.²¹ In some cases, however, ‘the contract was so desirable that the bey’s farmers would let their concessions to sub-contracted cultivators, in return for the [jabri], plus a supplementary advance’.²² Finally, ‘azl azib, ‘estate ‘azls’, were allocated to state functionaries in lieu of salary, and farmed by them as they saw fit.

Peasants were free to move, leave an estate between seasons and quit an employer, but ‘azl land could also be occupied over long periods by generations of the same population. Where settlement was stable, the right to live on and work the land was understood to be heritable to the male and female children of tenants, and only in the case of a tenant dying without heirs (either offspring or collateral kin) did the attribution of the land in question revert to the state. The crucial legal principle was that rights to land tenure and use were recognised through the labour by which land was fructified.

This was also the principle at work in a second main category of land tenure, later codified by the French as ‘arsh, or ‘tribal’ land. Often misunderstood as ‘collective’ property, what came to be known as ‘arsh lands were tracts of cultivable or pasture lands to which access was recognised as hereditary in given lineage groups, within which each family also had inherited rights to given plots (though these were never physically enclosed). The land was worked communally, but held individually, to the benefit of each individual family group. How much land a family could access – and correspondingly, how much of its produce they would receive – depended on their capacity to work the land, measured by the number of male labourers they could deploy: this notwithstanding the essential contribution of female labour to the group’s agricultural and domestic economy. The right to land was actually, therefore, a ‘right to labour’, which could not, in the case of ‘arsh land, be inherited by women, ‘who cannot hold the plough’.²³ A gender division of labour rights thus played a crucial part in the fortunes of family groups, where an absence of male children could mean the reduction of a household to dependence.

Just as a degree of individualised family property rights entered into the ‘communal’ ‘arsh system, conversely, a degree of communal rights attached to the third type of land tenure, milk or ‘possessed’ land, usually translated somewhat misleadingly as ‘private’ property. Again, the principle in most of the region – aside from the milk properties of notable landowning families – was that the degree of labour applied by the individual or family group determined the degree of individual rights as against community obligations in owning and disposing of property. Milk property, delimited by physical walls or hedges, was recognised as the ‘private’ property of families, who either held actual notarial titles to them or (just as legitimately) were said to hold titles derived from land grants made to vaguely recalled ancestors by undetermined past sovereigns. But while, for example, fruit trees planted and maintained by the family, land cleared and planted entirely by them or a house constructed by them could, in theory, be freely sold, cereal crop-lands or old olive trees, maintained at least in part by communal labour, could not. All members of the extended family, the lineage group and the locally defined community – including its members who might happen to be absent at the time, and for however long a time – exercised a right of pre-emption (shuf‘a) over any such sale, preventing the intrusion of outsiders and holding members of the community to account for the management of their supposedly ‘private’ holdings. ‘The power of [communal] association and of a community of interests’,²⁴ based in these fundamental ecological relations of land, labour and social reproduction, was thus extremely great, a kind of ‘tyranny of the community’ that created both durable ties of social solidarity and meticulously policed codes of behaviour and responsibility.

Nowhere, perhaps, was this truer than in Kabylia, the area where milk property was most common. Conventionally divided into ‘Greater’ and ‘Lesser’, or ‘upper’ and ‘lower’ regions, Kabylia (bled al-qaba’il, the ‘land of the tribes’, as seen from the point of view of town-dwelling lowlanders in Algiers and Constantine) is defined by mountain regions of the coast, from Algiers to Bejaïa and then eastwards to Jijel and Collo, and of the interior – the Djurdjura massif, whose highest peak, Lalla Khadidja, reaches 2,308 metres, and the Biban and Babors. ‘Greater’ Kabylia, on either side of the Sebaou valley, between the coast and the south face of the Djurdjura, is bounded to the south and east by the deep valley of the Soummam, which flows east and north to the sea at Bejaïa. Beyond this line to the south and east, the Biban and Babors make up the more extensive, but topographically less imposing, ‘Lesser Kabylia’.

Here, an especially densely settled population speaking the Berber thaqvaylith dialect built tightly packed villages on the hilltops and cultivated the well-watered slopes (between 700 and 1,000 mm of rainfall annually on the highest peaks) and valleys below them.

Cereals were grown in the river valleys and on small parcels of land flat enough for ploughing, but as in the Aurès, in Kabylia the peasantry invested especially in arboriculture and combined cultivation with shepherding. Olive oil and beeswax were regional exports. Artisanal production, too, in ironwork and jewellery as well as ceramics and textiles for domestic use, was an important part of the local economy, which, however, was frequently unable to absorb the energies of the whole population. Already in Ottoman times, Kabylia exported part of its male workforce as seasonal emigrants to the plains and cities. A densely populated area of fragile resources, Kabylia illustrates perhaps most famously the extent to which daily life, in such circumstances, can be bound around with densely elaborate symbolic codes governing land rights and the gender relations that underwrite family and lineage ties, procreation and the maintenance of a family’s access to resources and livelihood.²⁵

Figure 1.4

Kabylia. Foothills of the Djurdjura, looking south in the Sebaou valley near Tizi Ouzou. Note the clusters of housing on hilltops.

(Author’s photograph)

A final area to consider is the coastal plain, especially Algiers’ immediate hinterland, the Mitidja. An alluvial depression between the Atlas and the Algiers Sahel (the range of hills from which the Ottoman city faced the sea), the Mitidja formed a major part of the dar al-sultan, ‘the sultan’s estate’, a district governed and taxed directly from the capital. An enduring colonial myth held that on the arrival of the French, the Mitidja was nothing but a malarial swamp, and indeed, many early colonists died in its low-lying and badly drained western and eastern extremities.²⁶ It seems clear that the economic and political crisis of the Regency in the early nineteenth century had a deleterious effect on the area, which William Shaler, the American consul, described as having become ‘a perfect desert, without inhabitants or culture’ at the end of the Ottoman period. Shaler’s description is clearly an exaggeration for literary effect, framed as an indictment of the Regency’s ‘barbarous despotism’ – he also considered the Mitidja to be ‘probably one of the most valuable expanses of territory, its climate, position, and the fertility of its soil considered, that exists on our globe’.²⁷ A French reconnaissance mission in the plain, immediately after the fall of Algiers, reported that while part of the land was ‘arid and little farmed’, elsewhere it seemed ‘charming, well-cultivated and displaying extraordinary vegetation … The site … is one of the most beautiful one could see … a place of true delight’.²⁸ A century earlier, Thomas Shaw had described the plain as

For most of the Ottoman period, the Mitidja was intensively cultivated and highly desirable land; private gardens, orchards and country estates here grew fruits, vegetables, cereals and tobacco, pulses, herbs, maize, potatoes, chickpeas, melons, squashes, aubergines, cucumbers, peppers, tomatoes and grapes. The wine of Algiers, according to Shaw, could well compete with those of Spain and Portugal.³⁰ Enclosed estates (hawsh) or ‘gardens’ (jnan), either state-owned and let to tenants or privately held by the Regency’s elite, were worked by servile labourers, by immigrant workers from Kabylia and by the local Arab populations, who also held their own ‘arsh lands in the region recognised as watans (‘homelands’). While many such holdings were relatively modest, there were also great estates: Jnan al-Dey, a property of the last ruler of Algiers, Husayn Dey, comprised almost 30,000 square metres of orchards, vegetable gardens and vineyards.³¹

Across the country, the division of land tenure and access to it supported a heavily stratified society. Within the landowning category, holdings varied from parcels of less than one hectare up to groups of estates held by a single great family totalling more than a thousand hectares. Patrician, urban notable (beldi or hadri) families owned medium to large estates; the rurally based ‘lay’ aristocracy (jawad) and religious leaders (mrabtin) often owned more; the Regency’s governing elite, in their own right or through the distributive powers of the state, occupied the top of the pyramid. But while supporting a highly differentiated social hierarchy, the ecological balance of Ottoman Algeria seems generally to have provided the peasantry with a ‘decent’ standard of living; most people laboured freely and did not usually starve. Even the life of the lowest on the scale of free labourers, the proletarian, landless sharecropper or shepherd, was ‘less miserable than has often been supposed, precarious only in times of crisis’. Furthermore, ‘they do not appear to be very numerous relative to other cultivators’, that is, the independent tenant farmers and semi-nomadic, semi-sedentary occupiers of ‘azl or ‘arsh lands.³²

But as we have seen, the equilibrium of careful land use with climate and resources was fragile, and ‘times of crisis’ could be frequent, as could natural disasters like the earthquakes that struck Algiers in 1632, 1676, 1716 and 1755, wrecked Oran in 1790 and destroyed the town of Blida in 1825. After the regional demographic disaster of the Black Death in 1348, plague had returned frequently to Algeria. In 1654–57 it reportedly killed one-third of the population of Algiers.³³ In 1660–65, and again at the end of the seventeenth century, it combined with drought and locust invasion. It broke out again in 1739–42, 1752–53, 1756, 1784–87 and in the early nineteenth century. Drought or storms, producing poor harvests, especially when combined with locusts or increased taxation to fund regional wars (fought against both Fez and Tunis in the eighteenth century), or the stockpiling of grain for export by the authorities and their merchant intermediaries, resulted in frequent periods of famine.³⁴ An eighteenth-century French consul at Algiers wrote of a country suffering ‘all at once, plague, earthquake, and war’.³⁵ Life on the land was not wildly anarchic, nor despotically crushed, nor was it always precarious; for many, it was secure and sufficient, for some it was highly profitable. But it was often fragile, usually hard, and almost always frugal: ‘a home-oriented economy, to which the social unit turned for the satisfaction of its vital needs … Clothes were plain and food was simple, with coffee and tea considered luxuries.’³⁶ The systems of shared responsibility, recognised rights and common interests that provided for social cohesion and resilience in the face of frequent hardship came with sharp divisions of privilege and distinctions of status – by wealth, ethnic or religious affiliation, age and gender – and sometimes high degrees of dependence and exploitation.

‘Islamic’ Cities

A major factor of social distinction and a major element in the control and exploitation of the countryside, of course, was the town. Only a tiny minority of the overall population lived in urban centres – at the end of the Ottoman period, Algiers had perhaps 50,000 inhabitants: Constantine, capital of the east, 25,000, and Tlemcen, historic centre of the west though never the seat of Ottoman government there, 20,000. The regional government centres of Medea and Mascara were ‘very small towns’.³⁷ Nonetheless, for a variety of reasons urban life was considered ‘normative’. For an older school of scholarship, much of it founded on colonial-era writings about French North Africa, this ‘normativity’ of urban life was based on essentialist ideas about ‘the Islamic city’, an ahistorical notion of the centrality and defining features of urban living in a reified conception of ‘Islamic civilisation’.³⁸ City-dwellers in Ottoman Algeria certainly considered themselves, by definition, as more ‘civilised’, their learning and practice of Islam more orthodox, their lifestyles and economic activities more noble, than those of jabaylis, ‘mountain folk’, as the arabophone citizens of the caravan entrepôt, garrison town and vast palm-growing oasis of Biskra referred to shawi-speaking peasants from the Aurès to their north, or berranis, as rural people from the interior ‘open land’ were known in Algiers.

But this ‘civic ideology’³⁹ of distinction, though important for understanding the culture of city life and durable attitudes towards ‘the countryside’, should not mislead us into thinking that the city had an especially distinct kind of history.⁴⁰ The functions of cities, as concentrated centres of artisanal production, market exchange, legal arbitration, learning, political and religious authority, all existed in the countryside too, in locales of specialised handicrafts, at weekly markets, in customary deliberative assemblies of adult males (jama‘as, or tajma‘ats in berberophone areas), the person of the hakem or qa’id (district or local governor), the shaykh of a tribal fraction or the custodian (muqaddam) of a saint’s shrine. Population size is also no guide; parts of mountainous Kabylia could be so densely populated as to qualify as ‘urban’ areas,⁴¹ while some self-regarding ‘cities’ of the Ottoman period were dismissed by European observers as hardly more than villages. What was important about cities, as Lucette Valensi pointed out in trying to distinguish ‘cities’ from ‘towns’, was the perception of their distinctiveness: ‘Life-style – urbanity or rusticity – established the boundary … more than did the function of each entity’.⁴² Or rather, beliefs about ‘urbanity’ and ‘rurality’, defining perceptions of both, framed the ways in which people behaved and saw themselves, in a social as well as spatial hierarchy, disguising the fact that ‘city’ and ‘country’ were really only points of greater or lesser intensity on a continuum of social and economic functions. The distinctiveness of the city was more symbolic and ideological than practical and functional. If we understand this, it becomes easier to grasp the significance of the ‘mere villages’ that were many of Ottoman Algeria’s urban centres, and to see the relation between city and countryside as one of overlap and interdependence rather than – as city-dwelling writers often expressed it – in terms of vulnerable citadels of civilisation surrounded by seas of tribal ignorance and hostility.

Figure 1.5

Tlemcen, interior of the great mosque (eleventh–twelfth centuries) (K. Bouayad).

The practical as well as symbolic functions of cities were, of course, important, as intensive market centres, and as bases of the state’s coercive power and the dispensation of its law. The central government and that of Algeria’s three provinces (beyliks – also a term used for ‘the state’ in general) were based on regional urban centres. Algiers, as the Regency’s capital, developed an entirely new significance on the basis of its ancient foundations, as did Constantine, the ancient capital of Numidia in Roman times that became centre of the beylik of the east.⁴³ The smaller, central beylik of Titteri, south of Algiers, was centred on the town of Medea perched in the Atlas, and the capital of the beylik of the west moved from Mazouna to Mascara before the re-conquest of Oran from the Spanish in 1791.⁴⁴

But urban-centred power was impotent without effective means of transmission to the countryside through rurally based intermediaries. Hakems, qa’ids and shaykhs kept local order and represented local interests to distant authorities. Auxiliary tribal groups, so-called makhzen (government, literally ‘treasury’) tribes, furnished cavalry to the beylik in exchange for exemption from the taxes they helped raise from their ra‘aya (literally ‘flock’ or ‘subject’) neighbours. Men of the zwawa (an Arabic term from the Berber igawawen) population of upper Kabylia were recruited to the militia of Algiers and Constantine. Great power also lay with regional aristocrats like the Muqrani family. From the sixteenth century, the Muqranis were de facto rulers in the Medjana, an area of the high plains between Kabylia, the Biban and the Hodna mountains in the western Constantinois. They raised taxes and administered the region ‘as they saw fit’, in exchange for an annual tribute to Algiers and the guarantee of passage through their fiefdom for the Regency’s soldiers.⁴⁵ Conversely, some centres of state power, like Tizi Ouzou in the Sebaou valley, where the Ottomans established a burj (fort) which became the nucleus of an administrative settlement, or even Blida (bulayda, the ‘little city’) founded by the beylik at the southern limit of the Mitidja in 1535, hardly qualified as ‘cities’ in the eyes of patrician, hadri inhabitants of longer-established, more distinguished settlements, whether substantial commercial and princely towns like Bejaïa and Tlemcen, or smaller agglomerations like Mazouna, Nedroma or Ghardaïa. These apparently insubstantial places were steeped in a sense of urbane dignity and historical longevity that constituted a powerful civic ideology: the three last-named towns provide good examples to illustrate the point.

Mazouna, a market and textile-manufacturing town at the lower south-western edge of the Dahra mountains, an area of rich land farmed since antiquity and a point of exchange between the uplands, the coast and the Cheliff river valley, was described by the geographer al-Idrissi in the mid-twelfth century:

Agriculture, weaving and trade made the town live; what made it a city was not only its role as a centre for manufacture or the cultivation and exchange of produce but its self-perception as a centre of ‘civilisation’, founded particularly on its medieval madrasa (school of Islamic jurisprudence), ‘a high seat of culture and justice’ esteemed throughout the region and whose buildings were restored by the Ottoman authorities.⁴⁷ It was here that the great scholar Muhammad ibn Ali al-Sanusi, born in nearby Mostaghanem in 1791, and later founder of the Sanusiyya brotherhood that would become a major organisational force of Islam in North and West Africa before providing the first rulers of independent Libya, studied before travelling to Mecca and Medina.

Further west, Nedroma, nestled in the Traras mountains eighteen kilometres south of the coast and thirty kilometres east of the present Moroccan border, had a population in the early colonial period of only about 3,000 people, of whom 15 per cent were Jewish. The town (as the Berber ‘Falousen’, the almond trees) was mentioned by the geographer al-Ya‘qubi in the ninth century. Its mosque was built by the Almoravid dynasty in ca. 1090, forty years before that of Tlemcen. Contested between the rulers of Fez and Tlemcen since the thirteenth century, and between Morocco and the Ottomans from the mid-sixteenth century, the city was garrisoned from Algiers only in 1791, its internal politics being divided, meanwhile, between pro-Ottoman and pro-Moroccan factions. Its well-watered agricultural lands, livestock and grain markets, cotton-spinning, wool and pottery industries, its trading links to Fez and Tangier, its qasba (citadel), discreetly elegant courtyard houses and four-way division into distinct quarters were all important features of the town. But what made it a city was its citizens’ perception of their own historic dignity: the town’s internal government by a jama‘a assembling representatives of the principal notable families, whose president was invested as qa’id in command of the local militia; the endogamy of the elite families and careful preservation of their social capital through the education of their sons and placing them advantageously in commerce; an accent and dialect of Arabic distinctive to the town; the cultivation of the tradition of Arab-Andalusian music, with locally specific techniques of interpreting certain pieces of the repertoire.⁴⁸

Almost 600 kilometres south of Algiers and 200 beyond the Saharan Atlas, in the shebka (‘net’), a rock plateau in the desert crosscut with deep ravines, fenced off by a jealously guarded independence from the effective projection of coercion or arbitration by the Ottoman state, but forging a carefully negotiated relationship with it, was another group of cities based on an even stronger sense of historically distinctive community. The urbanity of the Mzab, a cluster of five cities in a valley of the same name whose river flowed only once every two or three years, was defined more by its particular religious and scholarly civilisation than by the exceptionally fragile local ecology, and the far-flung commercial activity, that gave the towns physical life. After the destruction in 909 of the Rustamid imamate at Tahert on the southern edge of the Tell, the Mzab was a refuge for its Berber-speaking Ibadi Muslims, adherents of a strict sect considering themselves true believers but denounced as heretics by the ruling Fatimid (Shi‘i) caliphs of the tenth century, and also considered such by the Sunni jurists of the Maliki school who later became dominant in North Africa. The principal town of the Mzab, Ghardaïa, was founded in 1053.⁴⁹ Date-palm cultivation was possible by irrigating the town’s gardens from wells, dug by hand down to the water table, and by a system of channels and basins designed to capture and distribute the occasional waters of the river, but life in the Mzab ‘depended on an influx of resources from the exterior’.⁵⁰

As a major Saharan market, Ghardaïa and its neighbours lived primarily from the profits of long-distance trade southward across the desert, with Mzabis organising and financing caravans manned and managed by Arabic-speaking nomadic Shaamba people. Increasingly, from the early modern period, Mzabi men also created an industrious diaspora of commerce and trades in the cities of the north. The constant preoccupation of the cities’ leaders and the goal of their economic activity was preservation of the community, identified from its origins as a threatened community of the truly faithful, bound by Ibadi doctrine – whose specificity lay more in community and political regulation than in theological doctrine or ritual practice⁵¹ – and sheltering in a true dar al-islam (‘domain of Islam’) from a world of corruption and unbelief (kufr). Institutions of Islamic learning and jurisprudence were central to community life. In the absence, since the destruction of Tahert, of a regularly invested imam, the cities were governed by jama‘as, councils of representatives of the city’s major lineage fractions (ashiras) who must be married and fathers, with established social positions and material fortune, presided over by the shaykh of the principal mosque and by the halqa (‘circle’) of ‘azzaba, the community’s ascetic religious scholars. All aspects of community life were governed by ‘agreements’ (ittifiqat), decreed by the shaykhs and ‘azzaba in consensus (ijma), which was considered binding on the entire community. Matters concerning women were regulated and policed by a group of women, the thimsirriden, consulted on such matters by the ‘azzaba. Their particular power derived from their responsibility for the ritual washing of bodies after death, a function they could refuse to perform for those considered as having transgressed the norms of accepted behaviour.⁵² A crucial element of community preservation lay in the strict ruling against women, and children up to a certain age, leaving the Mzab. Men frequently travelled north to work for long periods, remitting money to their city of origin for investment in trade and in date-palm groves (the latter, again, a symbolic commitment to the community). But throughout the Ottoman and colonial periods, women were forbidden to leave the valley. The community’s particularity was maintained, too, by the use of its Berber dialect as the maternal and domestic language, with Arabic (on mastery of which great value was placed) taught only from the time a child began to learn the Qur’an, ritual and religious sciences.

Mzabis who came north to the Regency’s capital entered a highly stratified but also mobile, diverse and polyglot society to which they were only one of several groups of immigrants. Indeed, the defining feature of Ottoman Algiers was perhaps the significance of its immigrants, who may have constituted a majority of the total population in the period 1580–1640: first their success as distinctly visible groups and then their absorption into the city, which combined their diverse origins into its own cosmopolitan identity.⁵³ From the late sixteenth century to the end of the seventeenth century, the ‘golden age’ of the privateer economy, Algiers’ wealthy elite was dominated by Andalusis and ‘uluj (singular ‘ilj: in Algiers, a term denoting a Christian captive converted to Islam, the ‘renegades’ of contemporary European writers).⁵⁴ Of a sample of forty-two inventories of the largest inheritances surviving from the period examined by the Algerian historian Lemnouar Merouche, sixteen are of ‘uluj and fifteen of Andalusi origin.

The Andalusis, refugees from the collapse of Muslim Spain who were often referred to in European sources as ‘Moors’, were an important and distinctive group of the Algerine elite. They are especially present in surviving sources as mu‘allims (master-craftsmen) and architects, merchants and manufacturers of silks and other textiles. The wealthiest combined the symbolic capital of religious learning as ‘ulama (singular ‘alim, a scholar of Islamic law and doctrine) with the practical enrichment of commerce, or took to the sea as privateer captains. From the sixteenth century, leading Andalusi families combined with existing Arab beldi lineages in marriage alliances, practising ‘a kind of caste endogamy, relaxed only occasionally, and reluctantly, and then only for the ruling individuals of the day’.⁵⁵

‘Uluj came from very different origins to these patrician elites, but their social promotion, after conversion and assimilation into the ruling society – not infrequently by adoption or marriage into the family of a former master – could be astonishing. Most of the superintendents of the treasury in the mid-sixteenth to mid-seventeenth centuries were converts. Ramdan ben ‘Abdallah, a convert and soldier in the Ottoman garrison, the freed slave and son-in-law of his former master, died in 1649 leaving in his will ‘a large estate, a magnificent house and four Christian slaves’ of his own.⁵⁶ The Regency’s governor during a succession crisis in 1556 was Hasan ‘Corso’, a Corsican convert, one of whose principal lieutenants was Ali ‘Sardo’, a Sardinian convert. In the crisis’ endgame, Hasan was executed on the orders of a pasha sent from Istanbul, who in turn was murdered by Hasan’s friend and ally, the governor of Tlemcen, Qa’id Yusuf, himself a convert originally from Calabria in southern Italy.⁵⁷ Another Calabrese, ‘Ilj Ali (‘Euldj Ali’ in contemporary European sources), captured in the 1530s, rose after his conversion from enslaved oarsman on an Algerine galley to privateer captain, Ottoman pasha of Tripoli in Libya, ruler of Algiers from 1568 to 1571, and Grand Admiral of the Ottoman fleet from 1571 to his death in 1587.⁵⁸ Less spectacular but perhaps more extraordinary was the family history of ‘the very learned mufti of Islam, Sidi Muhammad ben Sidi Ramdan ben Yusuf al-‘Ilj’ who in the 1630s was the Hanafi mufti (one of two chief judges and religious officials⁵⁹) in Algiers – the son of a senior ‘alim and grandson of a convert.⁶⁰ These are exceptional cases; more common was a degree of social and political promotion acquired through professional competence and personal ambition in a society that was largely meritocratic and where the fortunes of the volatile and complex early modern Mediterranean world brought together enterprising spirits from different horizons, capable of negotiating personal and familial destinies across the frontiers of social, political and religious divides.⁶¹ Marriage alliances were also contracted between wealthy and successful ‘uluj and families of Andalusi origin.

The Ottoman elite – as internally diverse as the empire itself, with the addition of ‘uluj from northern as well as southern Europe and the Mediterranean islands – monopolised political and military power, to which no ‘autochtonous’ Algerine beldi Arab or Berber family ever acceded directly. Many of them, particularly the ra’is (privateer captains), and the families of the deys (governors) themselves, amassed spectacular fortunes. The largest single inheritance attested from the Ottoman period is that of a daughter of one of the last Ottoman governors, Aisha bint Hasan Pasha, whose pearls and jewellery alone were valued at 730,428 pataques (97,390 Spanish piasters, SP); second in line comes the fortune of the last reigning dey, Husayn Pasha, at 591,159 pataques (78,821 SP).⁶² By comparison, the fortunes of leading merchants and high state officials registered at the same time, in the 1820s, generally fall in the range of 3,000–9,000 SP. Merchants in the Arab beldi elite sometimes joined their commercial fortunes to political as well as to more locally rooted, symbolic forms of power: Abd al-Rahman al-Barbri, an international merchant, came from a family whose marriage ties connected them both to the household of an Ottoman governor and to the descendants of Sidi Abd al-Rahman al-Tha‘alibi, a revered fourteenth-century scholar who had become Algiers’ principal patron saint.⁶³

At the other end of the scale were the city’s least willing immigrants: European Christian captives taken by privateers and Africans from south of the Sahara traded north by slave caravans. While those Christians who converted could enjoy rapid social promotion, there was also a significant differentiation in living standards within the enslaved population. European men, depending on their own background and the likelihood of raising substantial sums at home for their release, were more often held for ransom than enslaved for life. Women were very rarely ransomed, but however dependent as wives, concubines or domestic servants of their owners, their material position was often favourable compared to that of other captives, and their domestic position may have compared favourably to that which they could expect in their countries of origin. Indeed, many converted to Islam and married male converts, frequently from the same region of origin as themselves, creating a kind of expatriate community⁶⁴ – one in which women’s legal rights were better than those they would have enjoyed in early modern Europe. The conditions of captivity for Europeans, often dramatised in European accounts and particularly by the French and Spanish rédemptoristes, ‘redeeming’ religious orders like the Trinitarians devoted to raising funds and support for their missions to ransom Christians from the hands of ‘the Turk’, can hardly have been easy. But they cannot reasonably be compared to the mass chattel slavery endured by Africans in the Atlantic trade.

For one thing, ‘white slavery’ in North Africa was part of a long pattern of Mediterranean social and economic history in which people, like other resources, had been aggressively pursued and moved around the region since antiquity.⁶⁵ The Mediterranean slave economy was as much a part of life in Toulon and Naples as it was in Algiers, and eighteenth-century consular records, to take only a few examples, illustrate the extent to which, on both sides of the sea, those who had the misfortune to fall captive played parts in mutually understood diplomatic relations. Muslim slaves on French ships in Toulon were able to send a petition to Algiers to complain that they had ‘nowhere to perform their prayers in peace, and that their cemetery was not kept in proper order’. When reports reached Algiers of ill-treatment suffered by Algerian slaves in Naples, the dey in retaliation had the priests and officers of two captured Neapolitan galleys put to work in a chain-gang alongside ‘the murderers and criminals’ of their crews. On another occasion, the Neapolitans reassured the dey that their own Algerian galley-slaves were treated according to the rules – officers receiving particular consideration – and added, as from one rather superior sovereign to another, that ‘one should not be taken in’ by the complaints of ‘these sorts of people’ (on both sides). The king of Sicily on one occasion in 1756 declared himself uninterested in exchanging Algerian captives against his Neapolitan subjects held in Algiers, ‘but is still most vexed at the lack of vigilance of the latter, for allowing themselves to be captured’. After two French slaves escaped from Algiers aboard a French ship in 1776, the navy minister at Versailles warned the French consul that although the king approved the ‘wise conduct’ of the captain who had allowed them to go free on arrival in France, ‘you must ensure that such a thing does not happen again; His Majesty will even have the captain punished if the Dey demands it’. The consul was to reimburse the dey for his loss if required. Later the same year, a slave who had taken refuge on a French frigate was actually returned to the Algerians.⁶⁶

Some participants in this forced Mediterranean mobility were in fact not captives at all, but gave themselves up into servitude in Algiers as a preferable alternative to the rigours of their home countries’ military service or criminal justice. Spanish soldiers fled the grim presidio at Oran and sometimes ended up as janissaries at Algiers; professional people-traffickers assisted fugitive Europeans to cross from Italy and Sicily to North Africa as well as making the reverse journey.⁶⁷ In 1776, the French consul counted ‘some 180 fugitives from Oran and from different countries’ arriving in Algiers, against 131 captives brought in by sea.⁶⁸ In popular culture, rather than the abomination denounced by the redemptorists, Mediterranean captivity, however hard, was also simply one of life’s misfortunes, the stuff of adventure tales and love songs: as a seventeenth-century English lyricist put it, ‘In the midst of my sorrows, whilst others do mourn; ‘Tis the want of my Love that doth make me forlorn: … No torment like mine was, when I was a Slave: For the want of my Betty was worse than a Grave.’⁶⁹

Secondly, captives’ life chances varied enormously. Galley-slaves, like the Muslim prisoners and forced-labour convicts on European ships, endured much hardship:

The massive defensive works and harbour construction undertaken in sixteenth-century Algiers were also largely accomplished by slave labour. But from the seventeenth century onwards, once the fortifications were complete and the oared galley finally lost its pre-eminence in Mediterranean navies, the frequency of such experiences diminished. Other captives, as an English report of 1675 explained, were

It was customary in the eighteenth century for the French consul to provide bread, meat and wine to ‘the poor Christians, slaves at Algiers’ at Christmas.⁷² At the same time, European consuls, unable to hire reliable domestic servants from home, also routinely employed Christian slaves, affordably leased to them by the Regency, in their own households; though answerable to the discipline of the dey’s office, they came and went, exercised their professions, played cards, drank and got into fights as they might anywhere.⁷³ William Shaler, the American consul not generally inclined to mildness in his judgments of Ottoman ‘despotism’, reckoned in the 1820s that while slaves must certainly have endured the ‘occasional cruelty and hardship … inseparable from the unprotected situation of captives of any description’, on the whole

Some, indeed, did not leave, though free to do so. Baba Ali, dey from 1754 to 1766, had as his valet before his election a Sicilian boy whom he freed on acceding to the throne. Having initially intended to embark ‘for a Christian land’, the freedman instead opted to convert a few days after his manumission and remained in his old master’s service, having been promised an advantageous post in the administration. ‘The young man is very sensible’, commented the French consul, ‘and was able, while in the dey’s service, to obtain the freedom of several poor Neapolitan slaves who sought his help’.⁷⁵ Both women and men belonging to important officials are recorded as having left substantial sums in property – between 1,500 and 5,300 pataques – held in their own right at their deaths. A slave might be sold, at around the same time, for 250–400 pataques. In general, women captives seem to have held larger personal fortunes of this kind than their male counterparts.

By far the greater part of the servile and freed population, however, were among the poorest in society, as reckoned by the records of inventories of property at death: most left fewer than 100 pataques. The smallest such property we know of was that of a black slave whose worldly belongings were valued at three-quarters of a pataque. Black African slaves, frequently Muslims or at least nominally converted to Islam en route across the Sahara, could gain their liberty neither by conversion nor by ransom, though manumission was frequent, giving rise to a generally poor, low-status, free black urban population. Such freed African slaves tended to remain tied in close relations of dependence to their former masters’ families.⁷⁶ Black African Muslim captives were often no better off, and must generally have been in considerably worse positions, than their white Christian European counterparts.⁷⁷

Between the patrician elite at the top of the scale and the poorest slaves at the bottom, Algiers’ society displayed a wide spectrum of wealth and poverty, with a high degree of social mobility in both directions and many people exercising several professions in succession or at once. The ruling political class itself was highly volatile, with positions dependent on patronage, and offices, including the highest, changing hands frequently. An Ottoman ra’is could become rich and influential, be appointed to high office, fall from grace and find himself all but destitute, then experience a reversal of fortunes with a change of regime. Religious scholars holding positions as secretaries, imams (prayer-leaders) or court clerks could also be silk-merchants, bookbinders, saddlers or drapers; janissaries and minor privateer captains earned additional income as barbers, blacksmiths, shoemakers or owners of coffee shops.⁷⁸ Women, on the other hand, worked outside the domestic economy in only a few areas. Aside from the few women of saintly descent who could become guardians of mausoleums and recipients of substantial revenues from pilgrims or pious endowments (habus⁷⁹), women worked as midwives, food-stall holders, owners or employees of bathhouses, makers and sellers of candles or as prostitutes. The latter category, again, illustrates a diversity of fortunes. At the death of one woman registered as a prostitute, her property amounted to only 121 pataques. Another – the concubine of the mizwar, the agent responsible for policing and taxing prostitution in the city – left a sizeable fortune of 6,342 pataques.

The Jewish population – small overall, but ancient and significant especially in certain towns (perhaps 10 per cent of the populace, or 5,000 people, in Algiers at the end of the Ottoman period) – was itself stratified between a large, modest artisan class and a small elite of internationally connected merchant families. The latter, recently arrived associates of great trading houses based in Gibraltar, France, Livorno (Leghorn) and the eastern Mediterranean, were sometimes able to gain positions of considerable commercial and political importance close to the state elite – whose treasury lent to them, exploiting its own preconceptions, at 3 per cent interest.⁸⁰ The former, the great majority of the community, were descendants of immigrants who had come to North Africa in the ancient diaspora, of Berbers who had adopted Judaism in late antiquity or of fifteenth- and sixteenth-century Andalusi refugees. Long established in communities across the region, from the coast to the Sahara, and in the major nodal towns of trade, culture and manufacture from Ghardaia to Constantine and Nedroma, most continued long into the nineteenth and early twentieth centuries to occupy working-class positions, as couriers, porters or tailors’ assistants. Though sometimes living in distinct quarters and governed in civil matters by their own rabbinical authorities, there was often little to mark out the Jewish population from their Muslim neighbours: residential areas were not usually exclusive, and in many aspects of language, culture and profession they were indistinguishable from the other poorer classes of society.

Some immigrant groups, on the other hand, specialised in niche occupations which they dominated (without having a legally recognised monopoly⁸¹): Jijelis were bakers, Mzabis bath-house managers, Biskris porters and couriers. These groups of barranis, immigrants from the interior, were organised along with other professional groups in jama‘as (‘corporations’, known as ta’ifas in the Mashriq) under the authority of an appointed amin or head of the corporation, responsible for the quality of work in each trade, for taxation and levying fines. The chief amin, or amin al-umana, a hereditary post from the second half of the seventeenth century, met with the muhtasib (market inspector and censor), the qadi (judge) of the town and an appointed shaykh al-bled (chief of the municipality) in a kind of municipal council for the regulation of local and commercial affairs.

This mobile and cosmopolitan urban society gradually coalesced over the course of the eighteenth century, with the absorption of the descendants of immigrant groups – Turks and their children, referred to as kuluglis,⁸² Andalusis and ‘uluj – into a local urban elite within which foreign ‘ethnic’ identifications were no longer differentiated,⁸³ and the rise to prominence of barranis in the city’s craft and labour market. As slavery became, first, a state monopoly, and then marginal – with at most 500 captives in Algiers at the end of the 1780s – the role of the converts, pre-eminent in the sixteenth century, disappeared. Incomers from the countryside joined the ojaq (the corps of janissaries)⁸⁴ and became master-craftsmen and heads of corporations, roles earlier dominated by immigrants from Andalus or elsewhere.⁸⁵ In the final flourish of corsairing, during the Napoleonic wars in Europe, the chance of quick enrichment was opened up to barranis and their offspring: Ra’is Hamidu, the legendary admiral of the Algerian fleet killed in battle against the American navy in 1815, was the son of an Algiers artisan. While most migrants from the interior to the city remained poorly paid manual labourers, and some gained respectable positions in commerce, trades or crafts that carried with them the dignity of beldi status, a few managed to realise in fact the ideological image of the city as the place which, above all, could give access to riches, power and fame.

The Beylik and the World Beyond

Supported by rural production and capped by the division of power and privilege in the cities, the societies of Ottoman Algeria were shaped, beyond their local environments, by relationships to the state, to the larger community of Islam and to the wider worlds of Africa and the Atlantic, the Mediterranean and Europe. Algerians imagined and acted out their place in these overlapping contexts through origin stories that explained their ancestry, in the local landscape of sacred space and the practices of ritual and learning that symbolised the rootedness as well as the universality of Islam, and by their engagement in far-flung networks of exchange and mobility.

We have seen how the Regency came into being through a coalition of local urban politics and adventurous regional entrepreneurism in the context of Ottoman expansion; reference has also been made to the volatility and periodic crises of its government.⁸⁶ The unstable conjuncture of circumstances that saw its birth, and the frequent tumult of its high politics, however, should not disguise the basic continuities of Ottoman rule that gave the state substance over the three centuries of its existence. The synergy of local and immigrant interests that brought the state into being continued to enable it to function on various levels, from the security of land tenure and local markets that provided both production and taxation to the consolidation of notable family fortunes by marriage to high administrators or convert corsairs. The balance between local autonomy and recognition of imperial paramountcy varied over time, but both elements remained important to the end of the period. Khayr al-Din and his immediate successors ruled as semi-independent beylerbeys, governors general of North Africa, representing Ottoman power on the front line of its war with the Habsburgs which came to a stuttering end after the temporary destruction of the Ottoman fleet at Lepanto in 1571 and (conversely) the Ottoman recapture of Tunis in 1574. From the mid-1580s, pashas were named to govern Algiers for three-year terms, signalling more direct, but less durable, control by the Sublime Porte (the central Ottoman government). At the same time, privateering, though no longer a front in the early modern ‘world war’, entered a phase of massive expansion, increasing the power of the ta’ifat al-ra’is, the corporation of corsair captains, and the ojaq, who in 1659 overthrew the appointed pasha and attempted to establish their own military republic. The chronic instability of this system led in 1671 to the assumption of power, in turn, by the ra’is, who installed one of their own number as ruler, with the title of dey (from the Turkish deyi, ‘uncle’).

The rule of the deys, elected from among the divan or governing council that formed the city’s ruling oligarchy and confirmed in their positions by the Ottoman government in Istanbul, continued until the overthrow of the Regency in 1830. Algiers’ effective autonomy from the Porte was thus greatly increased, but the symbols of rulership in the deys’ Regency bore the marks both of locally legitimised and internationally recognised sovereignty and of attachment to the distant imperial centre. The dey was a recognised power in his own right in diplomatic terms: the treaty of peace with Britain signed in 1816 was formally ‘between His Majesty the King of the United Kingdom of Great Britain and Ireland, and His Most Serene Highness … Dey and Governor of the Warlike City and Kingdom of Algiers’.⁸⁷ But from 1520 until 1830, coins struck in Algiers consistently bore the name of the reigning Ottoman sultan and his title, ‘Lord of the Two Shores and Sovereign of the Two Seas’, marking the place of Algeria at the edge of a Eurasian empire that stretched from the Indian Ocean shores of Arabia through the Balkans to the western Mediterranean. The notion of belonging to the wider world of Ottoman sovereignty remained important, particularly for the Turkish-speaking military and political elite, but not only for them.⁸⁸ Into the middle of the twentieth century, Friday sermons in mosques in rural districts of eastern Algeria were still being said in the name of the last reigning Ottoman sultan.⁸⁹

At the local level, the due forms of sovereignty are clearly captured in an account of the investiture of a new dey by the military, political, scholarly and religious elites of Algiers in 1766. The chief officials, or nadi wuzara (‘circle of ministers’), ‘met and entrusted Mehmed Osman Pasha with rulership’, and the following day

The political history of the Regency has often been reduced to one of factional struggles within Algiers, between ra’is and ojaq, or ‘pure Turks’ and kuluglis, in which the ‘autochtonous’ population – whether beldi elite or peasant mass – plays no discernible role except to furnish extorted taxes. As we have seen, however, social dividing lines were not as clear, especially over the long term, as such views suppose, and the account quoted above, from an Algiers notable of a local Arab sharifian lineage, suggests the need to relativise the division between a ‘foreign’ ruling clique and the indigenous population. Other, more complex, hierarchical divisions were more significant. The beylik’s rule was never a matter of a few thousand ‘foreign’ janissaries holding millions of ‘natives’ in permanent subjection.

The effective limits of the state, and the extent to which its rule was recognised, varied over both time and space. When threatened by a hostile incursion from the Moroccan Tafilelt region in 1578, the people of the Gourara and neighbouring oases had no hesitation in seeking protection from the Ottomans in Algiers.⁹³ The inhabitants of the Mzab, on the other hand, ‘do not suffer their towns to be garrisoned, and the government of Algiers has never been able to make them submit’⁹⁴ according to a well-informed late eighteenth-century observer, but their trading privileges and freedom of worship were well established and recognised by the beylik, on whose stability their commercial prosperity and community survival depended. The Kabyles, depicted as almost perennially in revolt in the nineteenth-century account of De Grammont, certainly resisted effective Ottoman rule in the mountains and the upper Sebaou valley, where in the sixteenth century an autonomous statelet, the ‘kingdom of Kouko’, existed, founded by the Aït l-Qadi lineage whose regional influence continued long after their state submitted to the Regency in 1542.⁹⁵ A chain of forts in the Sebaou and Soummam valleys did little more than attempt to control communications in the region and enable some limited tax-raising. But equally, the bey of Constantine largely depended on zwawa recruits, from the igawawen tribes of the high central Djurdjura, for his army, and Kabyle labourers worked the estates and gardens of the Ottoman elite in Algiers and the Mitidja. The only massive, widespread rural revolts against the Regency’s government came in the first years of the nineteenth century, in a context of unprecedented general crisis.

The variable parameters of the state’s sovereignty are visible in economic and fiscal terms as well as in the extent of control over territory and population. Local and sectional interests within the elite frequently carried the day over any putative central sovereignty of the state. Control of foreign trade was sometimes considered a ‘state monopoly’ in contemporary outside sources.⁹⁶ In fact, it was a multi-tiered system of intermediary influence in which state officials, from the most modest tribal shaykh or local qa’id up to the khaznaji (treasurer), wakil kharaj (minister of naval and foreign affairs) or dey himself at Algiers, acted in his own as well as his administrative office’s interest.⁹⁷ In 1817, during an extreme crisis in Algiers brought on by widespread famine, the bey of Constantine was stockpiling grain ‘up to the ceilings’ of his warehouses at Bône, in anticipation of a high price for export.⁹⁸ The variability of both currency and taxation also illustrates the unevenness of integration under the Regency’s rule. From the early eighteenth century, the French Compagnie royale d’Afrique, which made substantial purchases of grain from Algeria throughout the century, used a specially clipped currency, by agreement with Algiers, in its transactions with Algerian partners, each coin being legal tender exclusively in the port or market for which it was intended – thus Collo, Bône (Annaba), La Calle (El Kala) and Constantine each had its own specially ‘adjusted’ piaster. Taxation too varied: hukr (groundrent) and ‘ashur, zakat or jabri (property taxes assessed on livestock or land under cultivation) were collected from areas, like the Dar al-Sultan around Algiers or the plains of Constantine, easily controlled from the city, but the authorities satisfied themselves with the more flexible lezma or gharama (head tax or ‘fine’) on less easily assessed populations, nomads or villagers in the mountains. The effectiveness of tax-raising declined with the friction of topography as well as that of distance from the centres of government, although into the mid-twentieth century, folk memories in areas like the Atlas mountains above Blida remembered the violence of Ottoman fiscal expeditions and still referred to lowlanders only half-jokingly as ‘Turks’.⁹⁹ The primacy of Algiers over the provinces, too, was symbolised by the dannush, a twice-yearly remittance of taxes from each of the three regions to the capital, and by the correlative confirmation of beys in post by the dey, but each province enjoyed a considerable degree of relative autonomy. Salah Bey, who remains a legendary ruler of Constantine in local historical memory, even attempted to assert the independence of his, most fertile and wealthy, province from the suzerainty of Algiers at the end of the eighteenth century.¹⁰⁰

As the state functioned between claims to imperial sovereignty and negotiations with local power and interests, so the cultural worlds of Algeria’s populations were defined by both universalist codes of meaning and belonging, and local hierarchical divisions. Most of the country’s people were Arabic-speaking Muslims who were ultimately descended from the Berber-speaking populations who had lived in North Africa since remote antiquity. Their own understandings of their ancestry and place in the world were governed by genealogical affiliations, from the extended family up to a tribal group of associated lineages, claiming descent from a single ancestor as aït …, awlad … or beni … (‘the children of … ’). Often, such associations were inherited by place of habitation rather than by actual descent, with successive migrations of new groups adhering to established lineages either by absorption or in degrees of dependence and clientship, but the genealogical idiom of shared ancestry served to anchor social solidarity in an imagined commonality of blood ties. Origin stories, preserved in oral and written literature, expressed the populations’ own histories on the basis of memories or myths of descent from ‘prestigious’ ancestors. The Arabic-speaking, pastoralist Awlad Na’il of the steppe and mountains in the Hodna region, for example, preserved the history of their ancestors in fragments of the epic of the Beni Hilal, Arab nomads whose migration to North Africa in the eleventh century became the basis for one of Arabic literature’s greatest romances.¹⁰¹ Stories of migration to, or of the founding of, settlements attributed their origins to early heroes of the Islamic conquests – as at Sidi Okba, founded at the desert’s edge south of the Aurès around the tomb of Uqba ibn Nafi, a companion of the Prophet and the first Muslim conqueror of the Maghrib – to virtuous warriors in the service of medieval caliphs, as at Nedroma, supposedly founded on the site of the tomb of one such martyr, or to Muslim saints (awliya, singular wali, literally one ‘near’ to God), as in the Gourara, where the foundation of ksur and the original irrigation of oasis gardens was frequently attributed to saints.¹⁰²

The saints, often termed mrabtin¹⁰³ (‘marabouts’) and their descendants, who inherited the baraka (charismatic power of ‘blessing’) of a saintly ancestor, provided the local infrastructure of Islamic learning and practice across the country, in towns as well as in the countryside. The tombs of such saints were often considered sanctuaries and repositories of spiritual power.

A mosque and centre of learning, sometimes with a library and hospice for students and pilgrims, might be attached to the mausoleum, the whole complex being termed a zawiya (plural zawaya). The most famous saints – Sidi al-Hawari (Lhouari) in Oran, Sidi Bu Madyan at Tlemcen, Sidi Abd al-Rahman al-Tha‘alibi in Algiers¹⁰⁴ – became, and remain, local patrons of the cities where their tombs were situated as well as regional scholarly or miracle-working celebrities to whose shrines pilgrims would come from far afield to pray, study or seek aid and intercession with God. Other, lesser saints’ shrines, domed and whitewashed qubbas over tombs or maqams marking a stopping place, were scattered across the countryside. They indicated the passing or the resting places of rjal al-bled, the ‘men of the land’, anchor points for the universalism of Islam and of learning in the local fabric of landscape and society. In the late eighteenth century, Husayn ibn Muhammad al-Warthilani, a scholar from eastern Kabylia, composed a massive rihla (travelogue) detailing his travels in search of the world’s ‘deserts and settled places, parched watering-holes and luxuriant gardens, cultivated villages and fortresses; the virtuous, distinguished and cultivated men of learning of all places; the shaykhs to whom knowledge has been given and the brothers who seek after truth … from West to East’. He describes the great mosques of Cairo in the same vein as his visit to the little coastal Kabyle town of Dellys, where he went ‘to make a ziyara [ritual visit] to [the tomb of] the shaykh, the pious wali, Sidi Ahmad ibn Umar, of whom I had heard in my childhood that he was among those to whom God had granted sanctity … ’¹⁰⁵ In this worldview, at once expansive and localised, there was no conflict between rural and urban, blessed saints and scholars of the law, the particular and the universal. From Kabyle awliya to the centres of learning and piety in Egypt and the Hijaz, the umma (community of the faithful) was rooted in the everyday.¹⁰⁶

Figure 1.6

Sanctity in the Sahara: the qubba of a saint, said to be Sidi Lhouari, near Timimoun (T. Sahrawi).

From the late eighteenth century especially, the custodian families of many local shrines and zawiyas became affiliated to one of the several Sufi (mystical) orders, or brotherhoods, of Islam whose networks spanned the Muslim world from West Africa to Southeast Asia, the most important in Algeria being the Rahmaniyya, Qadiriyya, Tijaniyya and Derqawa.¹⁰⁷ The orders, in Arabic tariqas (plural turuq) or ‘ways’, represented institutionalised hierarchies of learning and spiritual initiation, from disciple to shaykh, giving access to specific practices of communal and individual devotion. As well as memorising the Qur’an and studying the law, students (tolba) learned to recite the dhikr (‘remembrance’ of God) and the wird, a liturgy specific to the order, recited by permission of the shaykh. Such special learning, sometimes combined with esoteric or ecstatic practices leading to trance-states, sought attainment of haqiqa, the inner truth of things and especially of God, beyond (never instead of) fulfilment of shari‘a, the outward laws of religious observance. Both the immediacy and the universality of Islam were thus present and accessible across a whole spectrum of religious sophistication, from the everyday rationality of the ‘ordinary religion’ and cosmology of peasants to the rarefied gnosis of saintly shaykhs and their itinerant students. Abd al-Qadir ibn Muhyi al-Din, the great Algerian state-builder and leader of resistance to colonial conquest in the nineteenth century, who would in later life become one of the greatest ever mystical commentators on the Qur’an, drew his faith and his profound spiritual and philosophical learning from a succession of Sufi initiations and discipleships, beginning with his family’s affiliation to the Qadiriyya order, and proceeding through initiation into the Naqshbandiyya tariqa in Damascus while on pilgrimage in the 1820s, to affiliation with the Shadhiliyya way through his last master in Mecca in the 1860s.

This rich mix of linguistic and cultural diversity did not, of course, simply make a melting pot without friction and exclusion. But social divisions were not drawn straightforwardly along religious, cultural or linguistic lines, between clearly separate Muslims and Jews, foreigners and natives or Arabs and Berbers. Muslims and Jews living in cities like Nedroma or Constantine might see themselves as closer to each other than either was to their respective co-religionist in the countryside. Conversely, the distinction between groups of Arab or Berber ancestry was often elided in their shared belonging to Sunni Islam, the Maliki school of Islamic law and saint-mediated everyday religious practice. And, by contrast, Berber-speaking, sedentary Kabyle peasants, whose Islam was Sunni and Maliki and structured by mrabtin and tariqas, were as ‘foreign’ as nomadic Arabic-speaking pastoralists to Mzabi Muslims of the Ibadi rite, who were also Berbers, but who unlike the majority Sunni community recognised neither tariqas nor zawaya (although the Mzab had its share of saints and a densely sacred landscape of Ibadi cemeteries). Arabs and Berbers were less strictly defined ethnic groups than they were language communities, and language practice was often multilingual in contact zones between mountains and plains, at markets or in cities. It also shifted across generations, as Berber-speakers assimilated to Arabic-speaking society and (particularly in the Aurès) vice versa. Such processes were greatly accelerated with the population movements of the colonial period and have continued to the present.

Social status and belonging were codified by wealth and by local conceptions of race as well as by language and religion. Such hierarchical divisions within broader language and religious communities were at least as important as divisions between them. Social ranking by precedence of family – with whom it might be acceptable to exchange women by marriage, with whom it would be unthinkable – were most important. Arabic-speaking, nomadic tribal groups on the western plains and plateaux might be privileged makhzen soldiers, ‘aristocratic’ sharifs claiming descent from the Prophet or ‘subject’ ra‘aya. Berber-speakers in the Gourara were divided between isemghan, lower-status, darker-complexioned inhabitants of autochthonous or imported slave origin, and their lighter-skinned social superiors who supposedly originated in migrations of pastoralists from the north. In Kabylia, groups in certain settlements, or living near valley floors as workers on the agricultural lands of notables, were identified as ‘the descendants of slaves’ (eklan) and culturally coded as ‘black’ in distinction to ‘white’ imazighen or ‘free people’. The word imazighen became generalised to denote all ‘Berbers’ only in the second half of the twentieth century. The pale-complexioned Arab hadri (urban, urbane) elite of Tlemcen recalled their Andalusi origin against the darker-skinned Arab pastoralist inhabitants of the surrounding countryside. Such distinctions persist today. Arabic-speaking gnawa (‘people of Guinea’), black Africans originating in or south of the Sahara, were an identifiable professional and ethnic group in the cities of the north, often occupationally specialised and producing the distinctive style of music named after them. At Nedroma, where the gnawa community was very small, its members formed the nucleus of a specific tariqa which specialised in healing rituals and was called upon in need by the other communities.¹⁰⁸ In a country so strongly shaped by the intersection, crossing and re-crossing of intercontinental religious, cultural and political frontiers, the community ties of inclusion, and the boundaries marking exclusion, could for that very reason be both very localised and very sharp.

The Crisis, Recovery and End of the Regency

The Regency of Algiers was an early modern tributary state, broadly comparable in the nature of its prerogatives, and its ability to have them respected, to other such states elsewhere.¹⁰⁹ Privateering, which in contemporary polemics, colonial myth and some subsequent scholarship supposedly defined the exceptional nature of the state and provided its most crucial revenue, was marginal after the seventeenth century, replaced as a source of wealth by tribute payments, guaranteeing treaties of peace, from other seafaring states, customs and taxation, increased agricultural production, and trade, especially exports of wheat to Europe. Trading relations had existed, interrupted occasionally by famine and rural unrest, since the foundation of the Regency: French coral fishing and export rights were first established in 1547. In what Merouche has termed the ‘century of wheat’ after 1680, and particularly the second half of the eighteenth century up to the 1790s, the Regency enjoyed unprecedented general prosperity and stability, a period of which the quarter-century reign of Baba Mehmed Osman Pasha (1766–91) is emblematic.

The authorities in Algiers repeatedly claimed, in answer to instructions from Istanbul that the practice be definitively discontinued, that corsairing was essential to guarantee the salary of the ojaq and thus political stability, but this was no longer true in the eighteenth century. The scale and the profits of privateering, as well as the personal status and prestige of the ra’is, had declined. While the state took over from individual entrepreneurs as principal financier for expeditions, their importance in the revenues of the state and its elite was no more than ‘highly variable and uncertain’.¹¹⁰ The importance of corsairing, in which the Regency’s elite continued to invest, was rhetorical and symbolic, part of the regime’s ideology, an attachment to its origins and character as a ‘corsair state’ – like Malta¹¹¹ – in the seventeenth century’s ‘golden age’ of privateering, and to the image of ‘Algiers the Most Warlike’ as dar al-jihad, the ‘bastion of holy war’ – even when it was in fact at peace and profitably trading with most of the nations of Europe.¹¹² French ships in particular traded between Algiers, Marseille and other Mediterranean ports, carried Algerian pilgrims to Alexandria en route to Mecca and transported high officers of the Regency with their customary gifts between Algiers and Istanbul.¹¹³ Relations with France in the later eighteenth century were so cordial that in October 1777, Louis XVI personally ordered French naval vessels to assist in the salvage of an Algerian corsair shipwrecked near Perpignan. The officers and crew were received by the king’s lieutenant general in the province and returned to Algiers aboard a French ship carrying a certificate of their good treatment.¹¹⁴

But commercial dependence on Europeans and their fleets posed longer-term problems. From 1793, when Britain entered the European war against revolutionary France, and through the Napoleonic invasions of Spain, Italy and Egypt that followed, the Regency re-entered a period of geopolitical crisis on a scale not unlike that which had given it birth almost three centuries earlier. Now, however, the global distribution of power was radically changing, and the rules of the relationship between north and south, ‘the West’ and the states of Asia and Africa, were being rapidly rewritten.

Internally, too, the long period of stability up to 1791 was followed by several crises. From 1791 to 1817, eight deys ruled, their reigns begun and ended in a series of putsches – one, Muhammad Khaznaji, ruled for only fifteen days in 1815. Behind this political turmoil were economic and social stresses. Landowners and intermediaries had been greatly enriched by grain exports, a prosperity shared by the lower classes in the preceding years, but when food shortages returned and elites attempted to maintain high volumes of exports, unrest broke out in the countryside.

The worst crises came in the early nineteenth century. From 1803 to 1805 harvests failed and wheat prices increased tenfold, reaching a record at Algiers that meant fifty-six days’ labour were required for a construction worker to buy a measure of grain.¹¹⁵ Riots broke out in Algiers in June 1805. A prominent Jewish merchant, Naptali Busnac, seen as close to the dey and held responsible for shortages due to his involvement in the grain trade, was assassinated, and Ottoman Algiers’ first and only anti-Jewish pogrom killed perhaps one hundred people.¹¹⁶ In September, a revolt overthrew the dey. The crisis abated, but returned in 1814 when locusts devoured the harvest and was aggravated through the following five years: food shortages and plague, falling population and decreased agricultural production created a vicious cycle producing popular discontent and revolts that became widespread across the country, organised and legitimised by rurally based religious figures associated with the tariqas. In 1804, the Derqawa raised an anti-Ottoman insurrection in the west. Forces rallying to the Tijaniyya zawiya at Aïn Madhi resisted a siege by the beylik in 1820 and themselves besieged Mascara in 1827. Ferocious reprisals against rebels failed to stamp out revolt and increased the force of religious leaders’ denunciations of a corrupt and unjust government. A pattern of rural rebellion that would be repeated over the rest of the century against the French was already taking shape in the final years of the Regency.

So was the beylik in inevitable decline, its history doomed to end in unavoidable European conquest? Things were not so simple, and the Ottoman period cannot be seen merely as a ‘pre-colonial’ prelude to French rule. Not only would many of the foundations of society visible in this period endure long after 1830, structuring deep continuities in Algerian society that remain relevant today. A limited recovery of the state was even possible in 1817, when a new dey, Ali Khoja, came to power determined to break the cycle of instability at the summit of the state and its faltering control over the country. The memoirs of Ahmad Sharif al-Zahhar, naqib al-ashraf (‘syndic’ of the descendants of the Prophet, a notable dignitary) of Algiers, describe how, once properly invested, the new ruler dismissed his ministers, sparing some and executing or exiling others, and appointed in their place either the irrelevant and powerless or the exceptionally able, thus concentrating power in his own hands and those of trusted allies. ‘For the benefit of all the people and so that the dissensions created every day by the soldiers in the city should cease’, he secretly removed the treasury and his own residence from the Janina palace to the qasba (citadel) above the town. Leading mutineers among the janissaries were executed and ‘the fires of fitna (dissension) were extinguished’. A subsequent attempt by the ojaq to march on Algiers from the east was defeated by troops loyal to the dey. In a campaign of moralising zeal, attendance at communal prayer was to be enforced, ‘the suppression of fornication and of alcohol’ ordered and shari‘a penalties strictly applied in cases of contravention.¹¹⁷ This ‘revolution’, symbolised in Zahhar’s account by what to him were no doubt the properly puritanical strictures of good governance, was more practically an attempt to re-centre power in the state around the office of the dey, against the entrenched interests of the ojaq and with support from the marketplace, the urban elites and the Kabyle zwawa and kulugli elements of the army.

Ali Khoja’s attempted re-foundation of the beylik came shortly after the suppression of the janissaries in Tunis and Mehmed Ali’s massacre of the mamluks in Cairo in 1811, and anticipated the so-called Auspicious Event that broke the janissaries’ power in Istanbul in 1826. Such radical moves in these cases led to periods of sweeping reforms and the strengthening of the state. If Ali Khoja had such a revolutionary project, however, he did not live to see it through – reigning for less than a year, he died of plague in March 1818. A limited stabilisation continued under his chosen successor, Husayn Pasha, ‘a man of reason and piety, who respected the ‘ulama, the sharifs and the saints’,¹¹⁸ but he was unable either to contain revolt in the countryside or to strengthen the state against newly hostile outsiders. Internal troubles and commercial dependency coincided with a changing view of the Regency – and of the Arab and Muslim worlds generally – in post-Napoleonic Europe.

The restored European order that emerged from the French revolutionary wars and the Congress of Vienna saw itself as advancing in universal peace, rational government and ‘civilisation’, containing revolution with rationed doses of liberalism and preaching ‘liberty’ as its own watchword. It saw ‘piratical’ Algiers, where corsairing had briefly flourished again during the disruption of trade brought on by the European conflict, as a barbarous relic of a previous age.¹¹⁹ In April 1816, the English Lord Exmouth negotiated terms of peace with Algiers on behalf of Sardinia and Naples that included the free release, as British subjects, of Gibraltarian and Maltese captives, and indemnities to be paid by the Italian states for the release of Sardinians and Sicilians. This was still recognisable diplomacy in the old style, but European public opinion was unimpressed, and in August Exmouth returned, bombarded Algiers and demanded the abolition of ‘slavery’, the restitution of all Christian captives and the repayment of the indemnities – a properly firm action, to European eyes, on behalf of ‘civilised’ nations against a refractory ‘outlaw’ state. This, quite suddenly and without forewarning, was a new world, one to which the Regency’s elite was ill-suited, and to which they had neither the time nor the resources to adapt. Fourteen years later, a French army would descend with crushing weight of numbers and firepower upon Algiers ‘the Most Warlike’, Husayn Pasha would be sent into exile and Ottoman rule in Algeria was brought abruptly to an end.