Imperial Overreach

On 6 April 1869, a caravan of ninety-five men, 116 camels and thirty-six donkeys left Gafsa in what was then the Ottoman Regency of Tunisia. Their aim was to trade for grain at the market of Tébessa, conquered by the French in 1851 as one of their most southerly military outposts at the time. As soon as the caravan reached French territory, they divided into three smaller groups, each of which spent about ten days trading at local markets, unmolested. On 15 April, just before the first of these groups crossed the border back into the Regency, they were attacked by local tribesmen, mostly recruited among the Nemencha, and under the leadership of two of their three French-appointed leaders (qā‘ids) or their younger brothers. The caravan put up some resistance, and twenty-four of its twenty-seven members were killed.

The next day, the raiders took the bulk of their booty to the qā‘id in Tébessa, and on the way dropped by the local Arab Bureau to inform the two French officers stationed there of what had happened. That would have been the end of the affair, had not the father of one of the merchants who had been killed—Nani b. Chamouni, a Jewish trader from Gafsa—alerted the French-language settler press in Constantine. The latter jumped on the occasion of launching a campaign against what they saw as the excessive powers of the military in the Algerian countryside; both the Algerian leaders of the raid and the French officers in charge were therefore brought to—a much-publicized—trial. At stake here was not so much who would get the booty—much of which had already evaporated by the time the trial took place—as who had a right to engage in what kind of violence, where, and in whose name.

The raid took place during a tense moment in Algerian history. The Algerian East had been hit by successive droughts, locust invasions, and a cholera epidemic imported from France, at a time when the French army’s brutality and settler appropriation of communal land had destroyed crops and silos. By 1868, forty percent of the population of the Tébessa region had died of sickness or hunger (Taithe Reference Taithe2010: 119). It is comprehensible that survivors did not take kindly to the idea of foreign traders taking away the rare stocks of grain that still remained in the area, especially as they were escorted by tribesmen—the Hammama—who were known for their raids on Nemencha, and had indeed, in 1868, killed one of the Nemencha qā‘ids. In such a situation, the French government’s insistence on the importance of free trade made little sense either to locals or, it seems, to the two French officers in the Tébessa Arab Bureau.

The trial did not, however, investigate the dire economic situation or the colonial government’s failure to respond to it. Contemporaries rather saw it as part of the longstanding struggle between military and settler interests in the colony. Settlers had for some time chafed at military rule, condemning it as undemocratic, despotic, and contrary to progress—which for them entailed a “normalization” of legal relations in the colony to facilitate the expropriation of communal land. The two qā‘ids maintained that in attacking the caravan they had only followed French orders, as “nobody can ever act in this country unless under orders from the [French] authorities” (cited in Conseil de guerre 1870: 48). They also insisted that such orders to raid were quite ordinary. As Salah b. Rejeb—one of the indicted men, uncle to one qā‘id and father-in-law to the other—noted during the trial: “When we go raiding (razzier), it’s always on the [French] authorities’ command. We go with Colonel Flogny [the current incumbent de Sériziat’s immediate predecessor, 1862–1866] and Colonel Bonvalet” (cited in Conseil de guerre 1870: 23). This was exactly the kind of image of the French army in Algeria that settlers wanted to present to Parisian decision-makers and the broader French public.

De Sériziat, then in charge of the Tébessa Arab Bureau, freely admitted that he had initially ordered the raid and had only changed his mind—to his regret—after consulting with his superiors in Constantine via telegraph. As he further explained during the trial, he well knew that French officers and their qā‘ids needed to be careful not to be directly involved in the raids they ordered as part of their everyday duty: “We have, in our border posts, instructions that we have not drafted ourselves, and which recommend to anyone who is officially appointed not to get involved in retaliation, which means that, when they happen and the men want to cross the border, the appointed leader cannot follow and have to leave the command to their sons or brothers” (cited in Conseil de guerre 1870: 45).

Much of this caution was recent. Two years earlier, the regional commander in Constantine had authorized a raid by the Nemencha against the Hammama in Tunisia (Conseil de guerre 1870: 60), led by the then qā‘id Muhammad Gāba, who was then killed in action (one of his sons, al-Hafsī b. Muhammad Gāba, was one of the principal accused in the 1869 trial); and the Bey of Tunis had repeatedly found cause to call in the French ambassador to complain about frontier violations—often by high-ranking French officers themselves (Cutler Reference Cutler2017: 85).

Raiding Sovereignty

We can dismiss the defendants’ accounts as mere legal posturing, attempts to shift responsibility to others. But clearly they were, if not speaking the truth, at least providing a plausible defense based on assumptions shared by all those concerned, including the military court and the settler press. On the southern borders of empire, raiding was one technique of governance among others, and regularly conducted by French officers and French-appointed qā’ids; debates over whether these raids were authorized and by whom were part and parcel of the extension of imperial governance to its unruly margins, reminiscent of Benton’s description of early-modern maritime imperial worlds. But they were not legally recognized or indeed recognizable: the convenient category of “privateering,” so central to debates over the legitimacy of violence in the early modern maritime world, was quite simply not available.

This remained the case even though General Bugeaud, governor general of Algeria from 1840 to 1848, had proposed the term “razzia,” a Frenchified version of the Arabic ghazū, in an attempt to legitimize his brutal scorched-earth policy as “traditional native warfare” (Gallois Reference Gallois2017: 10). The term had been coined as a description of an extra-legal practice supposedly made necessary by the barbarity of “native custom,” as de Sériziat also claimed in his trial (cited in Conseil de guerre 1870: 51). Settlers balked at the extra-legal nature of much military practice, hence the trial, but it was part of a system of legal fragmentation in which certain areas and people within Algeria were legally constituted as existing beyond the law, as most famously through the indigénat (Mann Reference Mann2009)—thereby incidentally stretching the latter to encompass ever more legal exceptions.

In the Oued Mahouine case, it did so successfully: despite the publicity that accompanied the Tébessa trial, none of those indicted suffered any permanent consequences, although the court had no doubts over the illegality of raids of this kind. De Sériziat was fully acquitted and then made Officer of the Légion d’Honneur in 1872, while al-Hafsī b. Muhammad Gāba, the alleged killer of the Jewish merchant, was, after a few years of exile in Tunisia, appointed qā’id in 1882 (Establet Reference Establet1991: 35). The Tébessa region remained under military rule until 1912, and the “Oued Mahouine scandal” seems mostly to have paved the way for the invasion of Tunisia in 1881—a classic example of French-authorized raiding eventually turning into claims to territorial sovereignty.

If, from the French colonial point of view, razzias were classified as legitimately extra-legal, things stood rather differently with regard to local legal and political traditions, which might have informed the motivations of Hafsī b. Muhammad Gāba and his companions in addition to their evident political acumen and shrewd sensitivity to colonial power relations on the ground. For both within the Moroccan makhzan and the North African Ottoman Regencies, armed mobile expeditions had long been recognized not merely as necessary for the establishment of sovereignty, but as a form of sovereignty. As Dakhlia notes (1988), states in the Maghrib were expected to be mobile; in early nineteenth-century Morocco, armed tax-collection expeditions (mahallāt) did not represent but were the state: people knew that they belonged to it when they received a tent from the sultan as he was getting ready to depart (Laroui Reference Laroui2001[1977]: 92).

As Benton shows, this close connection between military predation and legitimate rule was not unique, but it relied on the ability to draw lines between authorized and unauthorized forms, lest mere “hordes of men” might declare themselves rightful sovereigns. In the eighteenth- and nineteenth-century Maghrib, mahallāt coexisted with harakāt, punitive expeditions against those who resisted taxation, and ghazawāt, raiding expeditions organized without the authorization of the sultan. Where mahallāt served to “show that the kingdom had a center and that this center was wherever the sultan happened to be” (Dakhlia Reference Dakhlia1988: 742) and harakāt could be described as a form of internal jihad, ghazawāt projected a different, horizontally negotiated, partial and decentralized image of sovereignty. Distinctions, however, were never cut in stone. Both the Ottomans and the Moroccan makhzan had few permanent troops and largely relied on auxiliaries recruited among local tribesmen—exactly those who in different circumstances might indulge in ghazawāt. Jihad relied on the existence of not-quite-Muslim-enough Muslims within the imperial realm, which was always a matter of perspective. The sultan’s sovereignty in this setting was thus largely a matter of authorizing certain groups to raid others.

In both cases, the key to power was not territorial control or centralization, but mobility, and the ability to direct and curtail that of others, as offers of protection went hand in hand with the ability to constitute oneself as a credible threat, in a pattern that the historian Benedetta Rossi (Reference Rossi, Quirk and Vigneswaran2015: 149) has aptly labeled “kinetocracy” and that is familiar also in the world described by Dua. Hafsī b. Muhammad Gāba and his companions might therefore be forgiven for their “imperial overreach”—which to their mind might quite simply have been a way of asserting their pre-eminence through whatever political language was available at the time. Yet the more general point remains: if the dividing line between lawful sovereignty and unlawful raiding was truly so fragile—a matter of religious opinion, easily-lost letters, oral messages, and hearsay—who was to decide which was which?

Questions of Law

When, thirty years after the Oued Mahouine trial, the French army command finally decided to push even further south, into the oasis belt of the Tuwāt, more than a thousand kilometers to the southwest of Tébessa, they noted with astonishment that there “was a qadi in every qsar [fortified settlement, pl. qsūr],” some of which had, at that time, barely two hundred inhabitants.Footnote ¹ Far from living in a lawless void, as northern scholars and colonial officers assumed, local inhabitants were able to draw on complex and sophisticated legal institutions, ranging from minor qadis in many qsūr, to the qādī al-jamā‘a (principal qadi) based in Tamantit or Tinilan, to a shūrā or majlis (council) of muftis, established in the second half of the eighteenth century in Tamantit, who met regularly and publicly to discuss difficult issues. And this is even without mentioning the many quranic school teachers, found in even the smallest of settlements; or the notaries, drafting written contracts even for the most insignificant of local transactions. By the second half of the seventeenth century, almost all of them were recruited and educated locally, in local quranic schools and zāwiyas (institutions of further Islamic education, Warscheid Reference Warscheid2017).

Whereas in the heartlands of Islam the costs of Islamic legal and educational institutions were met by princes or the wealthy, the Tuwāt was firmly and self-consciously situated “beyond the lands of the sultan.” Qadis, muftis, and zāwiyas here were funded primarily by local assemblies, councils of “notables” who (in local legal parlance) “stood in the place of the sultan,” until he should make his appearance. In the meantime, qadis were not only paid by assemblies, but they also relied on them for what little enforcement existed. Local assemblies, however, were not recognized by sharia, and assembly rule was generally cited, by Islamic scholars further north, as proof of religious ignorance: northern Islamic scholars, as Touati (Reference Touati1993: 98) writes, were “exasperated by the kind of local democracy” that reigned in places like the Tuwāt and that turned local assemblies into sources of legal authority—perhaps in much the same way as Benton’s lawyer Hoffman was horrified at the idea of “hordes of men” arrogating to themselves the privileges of sovereignty in revolutionary Latin America.

While it is easy to see, then, why local scholars might endorse the assemblies on which their survival depended, it is harder to understand why assemblies would go to quite so much trouble and expense to fund a legal system that had no place for them. The vexed issue of authorization touched on above might hold an answer. On one level, assemblies and scholars reciprocally authorized each other, and were clearly aware of their shared interests, however much they squabbled over legal details. Local traders needed qadis to authorize their commercial interactions in order to make them legible to their northern partners. Beyond this, if harakāt could be justified as internal jihad, a mission to spread sharia to the unruly corners of empire and beyond, then visible evidence that one already “had” the law might provide some protection.

At the very least, it provided a degree of self-confidence, which meant that assemblies could deal with other secular sovereign projects—the approaching French colonial army, for instance—with pretensions of equality. Hence the following letter, sent to the French commander in Laghouat in the 1890s by an “assembly of Tuwātī notables”: “We do not want to have any relations with you neither for buying nor for selling. Let everyone stay in their country; you where you are now, and we in ours. Although separated in this way, we would still like to receive and send good news to each other. Here are the names of the qsūr who agree with us and answer like us. [Here follow nine names.] These are those who think like us. Greetings.”Footnote ²

Booty into Property

But the need for law went even further, in a way that touches on the distinction, crucial for Benton’s paper, between booty and property. Assemblies defined themselves in their own internal documents not with reference to a “local community” of any kind, but with reference to property: they were the owners of vital pieces of collectively built and maintained infrastructure, mostly irrigation systems, but also walls and other forms of fortification. It was their right of property, and the benefits and responsibilities that came with it, that granted them the authority to issue local rules and regulations, to intervene in local residents’ lives to avoid strife, conflict, and the dispersal of local assets, to decree a rota of collective labor and hospitality to maintain these assets in a good state of repair, and to collect taxes in the “general interest.” And as Perl-Rosenthal notes in his contribution, this imbrication of private property and sovereignty made the latter inherently fragile.

This was brought home to the French when, in 1893, possibly the same year that the Tuwātī notables penned their letter, Muhammad b. Hamza, a well-known trader and perhaps equally well-known French spy, approached General Marmet, commander of the Medea subdivision. He suggested that instead of conquering the Sahara by force, the French could achieve territorial control by simply buying land and water in the region: “These acquisitions will all be made in my name…and it is only when we think that we have bought enough property in this way, that you will come and claim your property rights, using the title deeds that we will have given to you.”Footnote ³

Note here that Muhammad b. Hamza proposes to buy property rather than to just take it by force: it was rightful, not stolen, property that could translate into political influence. Although Marmet conceded that it was not a bad idea in principle, he advised against this proposal, on the grounds that it was too expensive and too time-consuming, but perhaps also because he had doubts over the possibility of converting property into the kind of territorial sovereignty the French craved.

If legitimate local political rule was a function of rightful property, then, as also pointed out by Perl-Rosenthal, the ability to constitute property as lawful, if necessary through “sheaves of documents,” was crucial. In the mobile and interconnected Saharan economies, this was, however, easier said than done. Oases relied for their survival on exchange and trade not only with northern traders, but also with local pastoralists, some of whom were less in tune with urban legal niceties. Local irrigation schemes were attractive investment schemes for wealthy outsiders of all ilk, and we can presume that, as in Muhammad b. Hamza’s suggestion, few questions were asked in such cases about the provenance of the wealth invested. Irrigation systems were dug and maintained by the enslaved, men and women raided from the south, in ways that might—or might not—have been legal, but bound oasis dwellers and war-like mobile pastoralists into close and intimate alliances. The legality of local sharecropping arrangements (khamassāt) was much debated throughout the Maghrib.

Worse, most oases paid some form of protection money to nomads, people who both held power unlawfully and lived through raiding, and could therefore, according to local scholars, be qualified as mustaghriqī al-dhimmah: people who had done so much wrong in their lives that they were “drowned in debt,” that is to say, all their wealth ought to be consumed in restitutions and compensation payments (Alazemi Reference Alazemi2026). Being defined as a mustaghriq al-dhimma had far-reaching consequences, as it spelled exclusion from all ordinary social intercourse, trade, and exchange, marriage most certainly, but also the simple sharing of a meal—perhaps like the “enemies of all mankind” central to much pirate-lore. In the “land beyond the sultan,” however, the mustaghriqī al-dhimma were not marginal outcasts, but held political power: unlawfully, certainly, but nonetheless effectively. Like Dua’s pirates whose kin are known to all, their invitations to dinner could not be refused, nor could their marriage proposals or investment schemes.

The only way in which the mustaghriqī al-dhimma and anybody associated with them could “return to God” was through a process called tawba or repentance, which involved handing over all one’s worldly goods to “the Muslim community.” Conflicts over who precisely represented the Muslim community—could act as an impromptu prize court in this fashion—were rife, of course, but overall the popularity of tawba resulted in what Berque (Reference Berque1970: 1349) called “the growth of the zāwiyas on the willing corpses of the badū [nomads].” In other words, the ability to legalize property was good business. Oases, through their investment in sharia, were machines that could convert booty into property, raiding into a respectable life, thereby producing non-state local forms of legality and partial sovereignty almost as a by-product. As such, they were perhaps akin to “households” in the Caribbean where wealth produced through raiding was transformed into respectable family lives (Benton and McGregor Reference Benton and McGregor2024)—through, much as was the case with local oases, enslaved and female labor.

Property into Territory

Muhammad b. Hamza had been quite right in pointing to the bottomless drain on resources that the French military conquest of the oases entailed, in particular as French territorial control hardly extended beyond oasis towns. Partly this was so because, far from simply replacing a former sovereign—a right that the French claimed had been granted to them by the conquest of Algiers from the Ottomans—they were attempting to establish a form of territorial control in an area where it had not existed prior to their arrival, or had existed only in a decentralized fashion within tiny oasis communities. But it was also because beyond (and even within) these oasis communities, an Islamically sound existence was not the only model of excellence that was available. Hassānī and Tuareg poetry and oral history, much like the Somali praise poetry mentioned by Dua, amply attest to values of a different kind, in which mustaghriqī al-dhimmah could lead fulfilling and prestigious lives, in ways that were less available to successful pirates in the Atlantic world. Many of these logics persisted well into the twentieth century and remained a tenacious nuisance to French colonial officers throughout.

Some of this might quite simply lead back to my initial observation that there is indeed a great difference between oceans and deserts: deserts, with their many ecological niches and possibilities, make it possible to live in the interstices of imperial and post-imperial shreds and patches in a way oceans do not, or do less; the audience Saharan raiders were playing to was not international (or New English or European), but regional, and raiders could hence live lives worth living without ever converting their booty into anything other than regional prestige, dependents, family, and herds. Finally, here too the particular materiality of the booty mattered: livestock are the only form of capital that really increases naturally, without the help of metropolitan banks. But this coexistence of different political forms also bears witness to the continued vitality and political autonomy of much of the Sahara even after the French colonial conquest was (nominally) completed in the 1930s.

Which brings us to our last point: the Sahara still has not been converted into territory, or at least not entirely so. Writing with the recent (2010s) Malian conflict in mind, Retaillé and Walther (Reference Retaillé and Walther2013: 597) hence argue that even today, instead of “weighing themselves down with territory and consuming their strength in the mastery of surface,” contemporary rebels and jihadis inhabit “very old political spaces,” those of the largely mobile empires that used to govern the Sahel zone from the ninth century onward (Reference Retaillé and Walther2013: 600, 606). Armed struggle in Mali is thus, in last resort, a clash between “two conceptions of space” (Reference Retaillé and Walther2013: 615), one territorial, one mobile—"a fabric full of holes…a tangle of strings”—which is why international military intervention, aiming to “restore” state sovereignty, has proven so ineffective. And this is perhaps where the most striking contrast lies: while Benton’s paper speaks of a world that is close to us in historical times, it is nonetheless mostly contained in the past. It points toward the messy genealogies of a (supposedly) much tidier present. In the southern Sahara, however, much like in present-day Somalia, this is how people today have to negotiate their lives on a day-to-day basis, creating opportunities for some, and endless burdens for others.