Ribât in Early Islamic Ifrîqiya: Another Islam from the Edge

Abstract

After a difficult conquest under the Umayyads, the eastern Maghreb or Ifrîqiya region was turned into the western borderland of the Abbasid Empire in the second half of the 8th century, and its governance was soon delegated to the Aghlabid Emirate (800–909). In this context, the Sahel (or Ifriqîyan coastline) quickly became a major centre of asceticism and pious collective retreat in places dedicated to ribât activities. This practice provided a framework for the life of devout people who kept a watchful eye on the Byzantine enemy while zealously performing their devotions. A genuine frontier society of religious men and devotees, ascetics and traditionalists arose in this burgeoning coastal fringe. Over the last two decades, this topic has given rise to a very rich historiography, notably produced by Tunisian researchers who have profoundly renewed our understanding. Based on these considerable achievements, the present contribution proposes to broaden the analysis in order to show how the rise of this movement of warrior piety, advocating an ideal of jihâd, must be related to a more global phenomenon, considered at the scale of the Abbasid Empire. Remaining in a comparative dimension, this article also proposes several approaches to the specific architecture of ribât sites, especially the place devoted to the community mosque.

1. Introduction1

In his well-known pioneering work, R. W. Bulliet masterfully highlighted how a so-called marginal region of the Arab Islamic Empire was gradually Islamicised in the first centuries after the conquest (Bulliet 1994). He describes with great insight the social and cultural dynamics that shaped the progressive assimilation of the Islamic religion by the regional elites and the fashioning of a new identity through the combination of local features and a universalist message. If many regions of the Arab Islamic world can lend themselves to the same historical reading, Ifrîqiya (the former Byzantine province of Africa, today’s Tunisia for its central part) stands out for the abundance of its sources, both textual and archaeological, as well as for a rich historiography that is today enjoying a full revival.

After a difficult conquest under the aegis of the Umayyad power, Ifrîqiya became, in the second half of the 8th century, the western borderland of the Abbasid Empire, whose governance was soon delegated to the Aghlabid Emirate (800–909). In this context, the Sahel, or the coastline which constitutes the maritime outlet of the regional political, religious and cultural metropolis at Kairouan, quickly became a major focus of attraction for the devotees of ribât.2 This Arabic term refers to a Koranic ideal of poverty and asceticism in the conduct of just war or jihâd; its adherents keep a watchful eye on the borders of the Muslim world while zealously performing their devotions (Chabbi 1995). Thus, in this rapidly evolving coastal fringe, a genuine frontier society emerged, made up of men of religion and devotees, jurists, traditionists and ascetics driven by a desire to fight the Byzantine enemy, and who resided in places specifically dedicated to their common activities.

This article aims to provide a synthesis of the prolific historiography (mainly in Arabic and French) produced in recent decades on ribât in Ifrîqiya. It also intends to place the phenomenon in a wider context, whereas until now it has been considered essentially from a regional and “national” perspective, and to propose new approaches to issues that have so far been overlooked or insufficiently examined.

2. Part I. A Historiography Undergoing a Major Renewal

Scholarship on the institution of ribât in Ifrîqiya has been favoured by the availability of relatively abundant and complementary sources: texts (above all biographical literature) on the one hand, and archaeological remains on the other. The two dimensions of this documentation have provided the basis for a scientific production that was sporadic throughout the 20th century and has become more constant and thorough in recent decades.

2.1. Milestones for a Historiography of the Institution of Ribât in Medieval Ifrîqiya

The text-based approach to the institution of ribât in Ifrîqiya is dependent mainly on exploitation of the rich material collected by Abû Bakr al-Mâlikî (d. ca. 1081) in his Riyâd al-nufûs, of which N. Amri provided a fine analysis while underlining its specific biases (Amri 2011). Other biographical works, as well as other types of sources, some geographical and some jurisprudential, complete this textual corpus, which can be regarded as well-furnished relative to the documentary conditions specific to other provinces of the medieval Islamic world. The historiography of ribât in Ifrîqiya can also rely on a substantial archaeological corpus, which provides an essential material counterpoint to the information drawn from the texts. Amongst the buildings that have been preserved, two monuments are of exceptional interest: the “Qasr al-Ribat”3 of Sousse and the “Qasr al-Kabir”4 of Monastir, which are systematically cited as examples. The other known buildings have not aroused the same scientific interest or are still unpublished. The state of archaeological knowledge therefore remains extremely uneven.

The first milestones in a historiography of the institution of ribât in Ifrîqiya were laid down long ago, during the colonial period, and were completed at the beginning of Tunisian independence; a detailed presentation is not provided here.5 After a long hiatus marked by a few scattered publications, it is only during the last twenty-five years that, from a historical and textual point of view, scholarship has experienced a noteworthy vitality. This is mainly due to Tunisian scholars, who are essentially working on a new, in-depth reading of medieval sources while taking advantage of the pre-existing archaeological literature. If we exclude R. M’rabet’s unpublished university dissertation (M’rabet 1988), the real historiographical turning point can be placed at the turn of the 20th to the 21st century, with the joint publication of the theses of N. Djelloul (Djelloul 1999) and M. Hassen (Hassen 1999, vol. I, pp. 58–62 and vol. II, pp. 735–43), as well as a paper by F. Mahfoudh (Mahfoudh 1999). In a broader framework, the simultaneous work by N. Amri on sanctity (Amri 2001) also falls within this trend. A few years later, A. El Bahi’s thesis (2004, in particular pp. 569–39) further marked out this field of research.

All of these studies pay marked attention to the impact of ribât activities not only on religious and social life, but also on settlement and the economy. The resulting renewal of our understanding is fundamental, but it must be noted that its scope hardly goes beyond the sphere of Tunisian academics. With a few exceptions, most of the production is in Arabic, which explains the limited influence abroad of such major works as those by N. Djelloul, M. Hassen or A. El Bahi.6

Similarly, but lacking any direct link with this movement specific to the Tunisian academic world, the question of ribât became topical again in European research during the same period. In France, this renewed interest is reflected especially in the work of C. Picard (Picard 2011); the study he published with A. Borrut (Picard and Borrut 2003) overlaps in several ways with the work of their Tunisian colleagues while broadening the topic to include the whole Mediterranean basin. In the Iberian Peninsula, the beginning of the new millennium was marked by major publications concerning the excavation of two ribât sites, Guardamar in Spain (Azuar Ruiz 2004) and Arrifana in Portugal (Varela Gomes and Varela Gomes 2004).7

Finally, in Morocco, the same period saw the discovery of Igiliz, another ribât site that gave birth to the Almohad movement (Van Staëvel and Fili 2006). The near-simultaneity of these different publications is noteworthy: it strongly suggests that this remarkable renewal of scholarship on medieval ribât should be considered at the scale of the western Mediterranean basin as a whole.8

2.2. The Institution of Ribât in Ifrîqiya According to Recent Historiography

Overall, the recent historiographical literature has provided a series of nuances or critiques of the way in which the first studies of the institution of ribât in Ifrîqiya, published throughout the 20th century, had dealt with this topic, while at the same time bringing new issues to the fore.9

Beyond a necessary re-evaluation of the overgeneralised use of the Arabic term “ribât”, which was erroneously applied to architectural forms (Picard and Borrut 2003, pp. 36, 51, 62; El Bahi 2004, pp. 322–23, 661–62), all recent studies have endeavoured to shed light on the diversity of functions that can be assigned to ribât sites. These roles were various and changing; some were concomitant, while others seem to have followed one another in time. The defensive value attached to most of the settlements, whether a true fortress or a mere watchtower, sometimes results in the identification of the most prominent ones as occasional refuges for the surrounding Muslim populations (Djelloul 1999, p. 199; El Bahi 2004, p. 597) and hence as possible nuclei of settlement (Djelloul 1999, p. 188; Picard and Borrut 2003, p. 50).

This protective function should not overshadow the religious dimension that is attached to any gathering of devotees. However, new interpretations have mainly focused on the economic dimension of these settlements. The evolution that can be traced from the fatwas of the 11th and 12th centuries suggests that they may have been stations on trade routes or warehouses (Djelloul 1999, pp. 88, 188–89, 199, 203; El Bahi 2004, p. 597). Far from defining establishments as fixed in a dual function, mixing defensive virtues and pious exercises, the institution of ribât in Ifrîqiya thus appears to be in constant evolution according to changing historical contexts. One recent achievement of scholarship is the ability to reconstruct this complex trajectory, which fluctuated with the major events that marked the history of Ifrîqiya in the early Islamic centuries. Following an initial security role, ribât establishments were affected by historical conditions and the political–military situation, which led to their inevitable reconversion.

Tunisian historians have thus embarked on the task of a more detailed periodisation of the history of ribât in Ifrîqiya and its embodiment in specific buildings, analysing its premises with the first official foundations under the initiative of the Abbasid caliphate and more precisely of its Muhallabid governors (771–795) in Monastir and Sousse (Djelloul 1999, pp. 185, 189; El Bahi 2004, pp. 579, 581), demonstrating for the Aghlabid period the quick decline in Byzantine maritime raids after the beginning of the invasion of Sicily and the corresponding loss of the defensive value of ribât sites (El Bahi 2004, pp. 577, 582–89), and finally contributing—probably in a decisive way—to the relativisation of the so-called “ribât crisis” that has long been proposed for the Fatimid period (El Bahi 2004, pp. 640, 721–22).

The daily life of these volunteers of the faith revolved entirely around a “dominant cultural paradigm” (Amri 2011, p. 367) consisting of a rigorous asceticism, initially carried out in tandem with guard duty, before rapidly evolving into a devotional practice free of any military activity. It is easy to reconstruct the ways in which ribât sites were frequented on the basis of textual evidence. Even if they welcomed fellow believers who sometimes hailed from distant geographical horizons (Djelloul 1999, p. 194; Hassen 2001, p. 155; El Bahi 2004, p. 622), most devotees came from the Ifrîqian Sahel itself (El Bahi 2004, p. 591). This explains the predilection for partial retreats during Ramadan, for example, or for seasonal stays of longer duration, while some chose to settle permanently (Djelloul 1999, pp. 75–76; Hassen 2001, p. 155; El Bahi 2004, p. 622). Tunisian scholars highlight the importance of the peregrine and “pastoral” (siyâha) vocation of the devotees who moved from one ribât place to another (Djelloul 1999, pp. 76, 192; Hassen 2001, p. 155; El Bahi 2004, pp. 622–23).

Within these activities, however, the handling of weapons seems of have been of limited importance: recent studies are unanimous in relativising the military efficiency of these ribât devotees and denouncing the primarily warlike vision of them that had prevailed previously (Djelloul 1999, pp. 196, 199; El Bahi 2004, pp. 591, 720). In fact, during the 9th century, martial duty seems to have been quickly limited to a symbolic, rather than an effective, watch under arms (Djelloul 1999, pp. 198–99; El Bahi 2004, pp. 335, 721).

Other qualities are put forward to celebrate the piety of these murâbitûn: austerity, voluntary or forced poverty, fasting, and the performance of supererogatory rites (prayers, vigils and nightly laments) (Djelloul 1999, pp. 192–93; El Bahi 2004, pp. 609–16). But these pious practices occasionally provoke reprobation on the grounds of being deplorable innovations (Djelloul 1999, p. 193). Another paradox of the institution of ribât also arises: while displaying a taste for asceticism, its adepts are sometimes rebuked for a lifestyle that benefits a little too ostentatiously from a very generous economic regime (Djelloul 1999, p. 195), for one of the crucial points of the new reading of ribât offered by Tunisian historians lies in the centrality of an economic dimension of the institution that previously had been poorly grasped. The recent historiographical trend has indeed highlighted the primordial importance of a landed property which, for the sentinels of Islam, constitutes an indisputable source of wealth and material comfort. By focusing on the relations of production and the methods of rational exploitation of environmental resources, these studies have profoundly renewed our vision of not only the institution, but also the economic history of the coastal region and the physical evolution of its landscapes.

Scholarly interest has focused, in particular, on himâ, the legal concept that governs the use of agricultural land and pastures set aside for the benefit of the murâbitûn. In this respect, the medieval sources add a slight documentary value, undoubtedly due to the contribution of a specific literature on himâ that has since been lost and is known only in bits and pieces (Djelloul 1999, pp. 84, 94, 201; El Bahi 2004, p. 653). The value of this ground-breaking enquiry lies in an interpretation that subtly articulates the modalities of implantation and territorialisation of ribât places along the coast on lands that were probably partly conceded by the central power, agricultural colonisation through the legal notion of himâ, and the exploitation of the natural resources of a hinterland that was dedicated first to feeding the devotees and then to supplying the markets due to the accumulation of surpluses, thus contributing to the economic development of the whole region (Djelloul 1999, p. 201; El Bahi 2004, pp. 641, 657). Unfortunately, we know very little about the management of these himâ-s, even if the information that can be gleaned shows that, over time, there was a shift from subsistence production to commercial agriculture, and from nourishing land to for-profit properties. The institution of ribât then turned into an agricultural exploitation in the strict sense of the term, accompanied by a mechanism of privative monopolisation of the productive himâ land. This land tenure system was strongly affected by the crisis that hit Ifrîqiya from the second half of the 11th century onwards (Hassen 2001, pp. 156–58).

The archaeological dimension of ribât, however, remains a weak spot in this renewed historiographical landscape. The paucity of reliable new archaeological data collected in the last half-century is notable, despite some work having been carried out,10 in particular by F. Bahri at Qasr ’Aliya (Bahri 2003) and the remarkable overall reflection proposed by C. Martínez Salvador on the origins of the buildings housing the volunteers of the faith in Ifrîqiya (Martínez Salvador 1998). But their scope remains undervalued to this day due to the compartmentalisation of historiographies.

3. Part II. The Formation of a Frontier Society

Current historiography thus offers a more complex, renewed and stimulating picture of the institution of ribât in Ifrîqiya and where it was exercised, cleansed of the stereotypes developed during the initial phase of research. This vision, produced above all by a generation of scholars highly versed in textual studies, nevertheless still presents a double problem of compartmentalisation. On the one hand, the international scientific community hardly takes into account the Tunisian bibliography, thus cutting itself off from a large part of the data now available. On the other hand, the horizon of this specific historiographical approach remains largely Tunisian-centred and barely reinserts the phenomenon of ribât in Ifrîqiya into a wider perspective, on the scale of the entire Islamic empire. It therefore seems appropriate to try to reconnect the Ifrîqiyan situation to broader historical perspectives in the following pages.

3.1. The Frontier as a Common Horizon

The regional tropism of Tunisian historiography can easily be explained by the nature of the sources used, the (relative) abundance of texts and archaeological remains that points the spotlight on the region of present-day Tunisia, and more specifically on the Kairouanese Sahel. However, this “national” approach to ribât cannot capture the general context in which it is embedded, which relates not only to a local situation, but also to social dynamics operating on the scale of the Islamic empire. In other words, a fine and nuanced contextualisation of the Ifrîqiyan case cannot be separated from the consideration of a more global phenomenon, of which it is clearly only a regional variation, without denying its particularities and originality. Far from being an eccentric feature, ribât is certainly one of the major modalities of the Islamisation process. Recent historiography has clearly underlined the close relationship that developed in 8th–9th century Ifrîqiya between the institution of ribât and the practice of jihâd on the edge of the Dâr al-islâm (Picard and Borrut 2003, pp. 56, 62; El Bahi 2004, pp. 615–17, 720). In medieval texts, the region is explicitly considered a political–military “frontier”, a thaghr. Put in context, this term inevitably conjures up the frontier zone par excellence (al-thughûr) of the Syrian–Anatolian borders disputed by Arabs and Byzantines. It was there, according to M. Bonner and his epigones, that the ideology of armed struggle on the frontier and of jihâd took its familiar form, with the appearance of an apologetic and biographical literature dealing specifically with this topic (Bonner 1996). The assimilation of ribât in Ifrîqiya to a style of life and action defined on the northern frontier of the Abbasid Near East was all the easier, as the enemy—the Byzantines—was the same.

One aspect of this phenomenon, which sees the institutionalisation of ribât in Ifrîqiya as the product of an official and centralised policy orchestrated by the Abbasid caliphate, was thoroughly examined towards the end of the 1990s. The official foundation of buildings intended to house the volunteers of the faith was accurately placed in the context of a global military strategy conceived at the scale of the whole Abbasid Empire during the last quarter of the 8th century. M. Bonner drew attention to the existence of strong governmental control over the joint provinces of Armenia, Azerbaijan and the Maghreb from the very end of the 770s via a network of governors and military chiefs under the supervision of Yahyâ b. Khâlid and the Barmakid family (Bonner 1996, pp. 77–78).11 This observation was extended to Central Asia by E. de La Vaissière, who underlined the perfect contemporaneity between the creation of ribât places at both ends of the Abbasid world: at the easternmost frontier, Fadl b. Yahyâ al-Barmakî, who was appointed governor of Khurâsân for one year in 794–795, was singled out for the large number of mosques and ribât places that he built; whereas at the other edge of the empire, Harthama b. A‘yân proceeded at exactly the same time (795–796) to found the Monastir fortress. Yet these figures all belonged to the Abbasid governmental apparatus and to the same social milieu, that of the Khurâsânian military commanders (quwwâd) of Baghdad (de La Vaissière 2008, pp. 74–75). Among these are prominent members of the Muhallabid family (771–795), who were involved in the difficult takeover of Ifrîqiya on behalf of the Abbasid caliphate (Bonner 1996, pp. 77–78, 147). Significantly, Harthama b. A‘yân had been entrusted in 793–794 by Hârûn al-Rashîd with the reconstruction of Tarsûs, a major ribât centre on the Arab–Byzantine border (Bonner 1996, p. 88).12 De La Vaissière can therefore rightly see in these events as “the beginning of a planned policy on the scale of the empire” (de La Vaissière 2008, p. 75).13

If the foundation of the first ribât sites in Ifrîqiya—official ones at any rate—tends to be better understood from now on within the framework of a global defensive strategy developed on the initiative of the Abbasid caliphate, the correlative development of Ifrîqiyan ascetic communities devoted to ribât is still implicitly seen as an endogenous phenomenon. Yet the study of the rise of the scholar, ascetic and warrior as a central character in the social and religious field of Ifrîqiya requires a broader vision which integrates the political and military situation that defined the struggle of the volunteers of the faith against the Byzantines on the Syrian–Anatolian border. M. Bonner clearly identified the process of construction of this ideal figure of the “scholar in arms” (Bonner 1996, pp. 107–34; Bonner 2006, pp. 97–101), a new and decisive actor in the frontier warfare, who is still only occasionally referred to as murâbit in the texts (Bonner 1996, pp. 108, 159, 165, 179).14 Indeed, the term ribât seems to be little used in the thughûr (Bosworth 1992, pp. 285–86; Eger 2015, p. 228),15 and the very nature of the places where it was practiced is not clear from the textual sources (Eger 2012, p. 437). Nevertheless, the notion was important to contemporaries: located on the borders of northern Syria, Tarsûs was undoubtedly one of the main points of this ribât activity on the frontier, as evidenced by a well-known quotation from Ibn Hawqal (Bosworth 1992, pp. 284–85; Eger 2012, p. 437). As for the connections between this thughûr region and the development of ribât activities in Ifrîqiya, they are tenuous but nonetheless evident. The Riyâd al-nufûs, for example, preserves the record of a scholar–recluse in Monastir during the 9th century, Abû ‘Amr Bashîr b. ‘Amrûs, who on his return from the Pilgrimage journeyed to Bilâd al-Shâm to stop at Tarsûs, where he joined a group of virtuous devotees (Mâlikî, I: 419). Other North African devotees are known to have been familiar with the region (Bosworth 1992, p. 282), such as Abû Bakr Muhammad b. Sa‘dûn al-Jazîrî al-Tamîmî (d. 955–956), who used to visit the ascetic retreats of north-western Syria and even took part in raids (Mâlikî, II: 409). The reference to Tarsûs was flattering and evocative, even on North African soil: Sousse is thus referred to on occasion as “[the] Tarsûs [of the Maghreb]” (Mâlikî, I: 486).

3.2. The Ifrîqiyan Version of the Ascetic Warrior

In this brief reminder of the situation in the Syrian–Anatolian thughûr, the similarities with Ifrîqiya obviously draw attention. It therefore seems highly relevant to reconnect these two occurrences of ribât activities, on the northern confines of the Abbasid Empire’s core on the one hand, and on the Ifrîqiyan coast on the other. In the latter region, the scholars in arms are described in terms quite similar to their Near Eastern counterparts. These ascetics form a new society that breaks free from the old bonds of solidarity. Almost exclusively male, this form of militant sociability promotes chastity, continence and renunciation of the pleasures of the flesh (Djelloul 1999, pp. 193–95; Hassen 2001, p. 155; El Bahi 2004, p. 624), even if this position of principle is often undermined in reality by the marital situation of many seasonal devotees (Djelloul 1999, p. 194; El Bahi 2004, pp. 624–25). In this context, murâbitât women are still the exception; their presence, however, is well noted by al-Bakrî in Monastir (Djelloul 1999, pp. 101, 194).

Likewise, if not more so than in the Near East, the fortresses of the Ifrîqiyan coastline seem to constitute, at least initially, the outlet for an aristocratic practice of frontier warfare, reconciling martial and pietistic tendencies with a high-ranking type of sociability. This was perhaps a way for its adepts to preserve a certain form of distinction at a time when the dynamics of social integration and conversion to islam were beginning to grow in Ifrîqiya. The social origins of the murâbitûn are sometimes presented as varied (Hassen 2001, p. 155), but some clues nevertheless point to the frequentation of ribât sites by members of the privileged social classes: it is enough to recall here the assiduity of Sahnûn (d. 854), who was also an eminent landowner, in his practice of occasional retreats during Ramadan. With its exacerbation of a warrior masculinity, military activity appears to be a practice of social distinction, demonstrated by the interest shown in cavalry (Djelloul 1999, pp. 196–97; El Bahi 2004, p. 593).

The re-contextualisation of the institution of ribât in Ifrîqiya within more global social and religious practices attested in particular on the Arab Byzantine–Syrian frontier also allows us to overcome the apparent contradiction, raised by Tunisian historiography, between the formal militarisation and the poor military efficiency of these devotees, portrayed as half warrior guardians and half ascetic practitioners. In fact, this effective weakness is of little importance when compared to the discourses and images conveyed by this warlike spirit and these martial attitudes. In the biographies provided by M. Bonner regarding the thughûr, none of the ascetic scholars appears to have been killed in battle (Bonner 1996, p. 158). The historical context specific to Ifrîqiya may partly explain the warrior ideals advocated by the adherents of ribât. During the last decades of the 8th century, the region was just beginning to recover—not without some jolts—from a century or so of unrest and chaos. The institution of ribât surely contributed then to channelling the actual or latent violence in the local society and to directing the mobilisation of the most aggressive actors against the Byzantine enemy, and not—symptomatically—against the Khârijite entities that had on several occasions threatened the hold of the two Near Eastern caliphates over the region.

As was the case on the thughûr, adherence to the ideals of ribât and jihâd was nourished in Ifrîqiya by a common imagery, that of the initial geste of the Arab conquest. A corpus of hadîth-s, some of which were forged during the second half of the 8th century, contributed to the elaboration of an ethos of the warrior ascetic, which summoned the memory of conquering heroes and aimed to perpetuate the precepts of primitive islam (El Bahi 2004, pp. 618–19; Amri 2011, p. 341). The social space of the frontier provides the opportunity to recreate the conditions for a particular mode of sociability and action, made up of pious companionship, community loyalty and personal effort, and martial and renunciant piety. The military weakness that can be glimpsed in the sources at our disposal thus appears to be counterbalanced, in Ifrîqiya as on the Syrian–Anatolian border, by a heroised piety: the community of the chosen ones is distinguished not only by a rigorous obedience to divine commandments, but also by a more assiduous practice of acts of worship, by means of a true culture of performance (Ehinger 2019, pp. 276, 280): fasts and supererogatory prayers, practices of reclusion, refusal to mingle with the social world, etc. Far from being limited to Ifrîqiyan or Syrian scholars in arms, these characteristics correspond much more broadly to the analysis proposed by T. Sizgorich of extreme movements of ascetic piety and violent militancy in the religious culture of the Late Antique Near East (Sizgorich 2009).

This comparison on a larger scale and the emphasis on similar practices, no longer in terms of “continuity” between Christianity and Islam but in terms of a shared culture, consequently, allows us to reassess the recurrent issue in historiography of the “influences” of Christian monasticism on the institution of ribât in Ifrîqiya. This question still depends on a reading that attempts to evaluate the “continuity” of practices, a position that is strengthened by the toponym “al-Monastîr”, which is attested in several parts of Ifrîqiya and is transparent in etymology. Recent Tunisian historiography discusses the numerous convergences and deep similarities, with regards to practices and beliefs, that it claims to find between Late Antique Christianity and nascent Islam: it sometimes sees in them hints of an undeniable continuity (Djelloul 1999, pp. 185–86, 193). However, the Arabic texts do not help to solve this problem, since they typically barely mention (and more often do not mention at all) the Christian presence, past or present. Be that as it may, the problem is perhaps simply poorly framed by historiography: for example, A. Nef’s recent reflections on the birth of the social world of Islam in the Mediterranean demonstrates the need to move away from a historiographical vision based on the dialectical categories of rupture and continuity (Nef 2021, pp. 57–63). It is therefore necessary to shift the issue to different ground. Even if the relationship between monasticism and ribât in Late Antique and Early Medieval Tunisia cannot be clarified in the current state of the documentation, one can nevertheless reason that any institutional vacuum invites a substitute, and that the residence of the murâbitûn recovers many functions and advantages that were formerly the prerogative of Christian monasteries, without suspicion of any direct “influence”. In other words, the practices associated with ribât do not necessarily directly continue a pre-Islamic monastic habitus. Based on the well-known hadîth of ‘Abd Allâh Ibn al-Mubârak (d. 797), “Every community has its monasticism; that of my community is jihâd in the way of God”, T. Sizgorich’s insightful analysis shows that, compared to the ambivalent figure of the monk, a man of prayer and action according to the vision offered by Muslim texts, early Islam developed its own version of militant piety embodied in the figure of the zâhid, the devout adept of renunciant piety (Sizgorich 2009, pp. 168–95; Sahner 2017). The same is most likely true in early Islamic Ifrîqiya.

3.3. The Importance of Land Tenure

As mentioned above, one of the major strengths of Tunisian historiography has been in highlighting the issue of land tenure, which is closely linked to the development of the institution of ribât in Ifrîqiya. This issue, too, undoubtedly deserves to be considered in a wider context. M. Bonner has clearly shown how important this dimension was on the Anatolian border, and how it contributed to the competition that emerged, starting in the last quarter of the 8th century, between two antagonistic social groups: professional soldiers on the one hand and warrior-like scholars and devotees on the other. The Abbasid caliphate, whose presence was strengthened by the transfer of the seat of government to Raqqa under Hârûn al-Rashîd and the creation of a new frontier province, the ‘awâsim, took care not to let either of these two groups run wild by maintaining its control over access to land resources in order to limit the possibility of creating large estates, and hence the formation of a warrior aristocracy (Bonner 1996, pp. 143, 152–53).16 This situation can undoubtedly be transposed to Ifrîqiya between the end of the 8th century and the Aghlabid 9th century. It is indeed quite remarkable that the Sahel region, where a good number of ribât sites were progressively concentrated, was so little marked by the presence of the regular army, the jund: the defence of the territory was in fact entrusted to the scholars in arms, who are preferentially established there (El Bahi 2004, pp. 594, 720). This was undoubtedly a conscious choice by the central power.

The general topic of land use in the Sahelian zone from the Arab conquest to the early Aghlabid period has been well elucidated in recent decades by Tunisian historiography. The region, abandoned by the Byzantine elites and emptied of its inhabitants, then marginalised during the first century of the Muslim presence, appears to have been subject to few attempts at colonisation due to the prevailing insecurity until well into the second half of the 8th century (Djelloul 1999, pp. 186, 200; El Bahi 2004, pp. 174–94, 548, 642–47; Hassen 2001, p. 156). Although the legal status of the lands thus abandoned is not easily grasped, it seems that these areas were considered conquered lands and thus assimilated to state property (El Bahi 2004, pp. 174–94, 209, 648; Hassen 2001, p. 156). According to this thesis, the recovery, cultivation and repopulation of these hitherto neglected or underexploited lands could not be entrusted to members of the jund, a dangerous social class whose turbulence and recurrent revolts against the central authority could only arouse fear within the government apparatus, and whose access to land in the hinterland of Kairouan could have led to their control of the land and their takeover of its resources. These public lands were therefore assigned to other actors, namely the murâbitûn—a way for the state, if not to keep control over them, at least to avoid the formation of an aristocracy of professional warriors.

The key point of this operation is how the land was granted, which, as we saw earlier, involved the imposition of himâ legal status on state property. Even if the conditions for the institutionalisation of these reserved areas on behalf of the first ribât sites are difficult to perceive in the particularly thin documentation that has come down to us (El Bahi 2004, p. 641), we do know that the procedure took a very official turn, since it was the chief judge of the entire province, i.e., the qâdî of Kairouan ‘Abd Allâh ibn Ghânim (in office from 788 to 806), who oversaw the founding of these estates (Djelloul 1999, p. 185; El Bahi 2004, pp. 578, 595, 647). It is significant that this official, who had been born in Ifrîqiya, studied partly in Iraq and Syria, perhaps in close contact with some of the warrior-ascetics who lived on the Arab–Byzantine border (Mâlikî, t. I p. 215). The colonisation of the coastline and the revivification of the land that accompanied it thus took place within the very specific context of the development of the murâbitûn movement (Hassen 2001, p. 156). The conquest of the coastal margins, which was now authorised by the gradual concentration of means and men and further fuelled by regular donations that reinforced the process, then began.

4. Part III. The Ribât Site: Building Context and Architectural Specificities

Aside from the texts, the other documentary aspect that has long been highlighted and explored by historiography is that of the material and architectural setting within which the activities of the murâbitûn in Ifrîqiya were carried out. This aspect has given rise to a number of early studies that today still remain largely authoritative; the renewal of our understanding in this area is much more modest than in the field of textual sources. In fact, the state of research on the architecture of ribât sites is still something of a trompe-l’oeil, with the “Qasr al-Ribat” of Sousse and the “Qasr al-Kabir” of Monastir dominating the foreground and relegating the other buildings, which have been much less studied, to the background.17 This part of the article aims to address two important issues directly involving these monuments. The first concerns the nature of their foundation, private or official: this lies at the heart of the ambiguous relationship between the central authority and the milieu of the armed devotees during the Aghlabid period in the 9th century. The second concerns the morphology of the ribât fortress; it deals less with the origins of its singular architecture, which is very characteristic of early Islamic Ifrîqiya and has already given rise to a great deal of discussion, than with one of its architectural features, namely the presence of a prayer room inside the building. The absence of a large amount of usable information for the 10th and 11th centuries leads us to focus on the Aghlabid period, the only one that is well documented on these two points.

4.1. The Modalities of Building Patronage

Ascetics and paragons of virtue on the one hand and scholars in arms on the other, the Ifrîqiyan murâbitûn were warriors in a category of their own, since they were not paid for military service by the central government. As in the case of the Syrian–Anatolian border, ribât thus appears from its emergence in Ifrîqiya as an institution totally distinct from the official military system (the jund); in other words, the physical framework of ribât was not a place of garrison for the regular army (El Bahi 2004, pp. 594, 618, 720). Like their fellow devotees stationed in the thughûr, the Ifrîqiyan sentinels of Islam cultivated a feeling of pious scrupulosity towards the lure of power and a similar partial rejection of state authority: biographical texts highlight their reluctance to mix with the ruling elite, even though the latter sought to win their favour by various means (pious visits, appointments to official functions) (Bonner 1996, p. 108; Djelloul 1999, p. 74; El Bahi 2004, p. 634). At first a military post supporting an imperial policy of defending the borders, the ribât site was subsequently converted, during the 9th century and as the Byzantine danger receded, into a spiritual bastion pitting a more or less intransigent Mâlikite Sunnism against state Hanafism (El Bahi 2004, pp. 611–14, 629).18 But this posture of disdain and defiance never led to open rebellion. On the contrary, the movement displayed a quietistic attitude and little inclination towards armed uprising: the absence of interference of the murâbitûn in the struggles between the jund and the Aghlabid state is particularly notable (El Bahi 2018, p. 335). During a good part of the 9th century, therefore, the Sahel region, and in particular its coastal fringe, constituted a place where consolidation of the state apparatus and the constitution of a religious counterpower based on the ideas of exemplarity and admonition were expressed in a competitive manner, without the professional army having a say in it.

It is obvious that, in such a context, the issue of initiative in architectural commissioning and building supervision appears to be crucial for understanding the nature of the relations between the central power and the milieu of the devotees. Unfortunately, no comparative data are available for the Near and Middle East, with the notable exception of E. de La Vaissière’s study of ribât in Central Asia. As C. E. Bosworth and then A. E. Eger have shown in their works (Bosworth 1992, pp. 285–86; Eger 2012, p. 437; Eger 2015, p. 228), there is indeed little information on the material expressions of ribât on the Arab–Byzantine frontier, as well as on the conditions that prevailed at the foundation of new buildings or settlements intended to accommodate the volunteers of the faith. The nature of the sources at our disposal for Aghlabid Ifrîqiya, on the other hand, offers some materials likely to help us in examining the strategies behind the foundation of such establishments. On the basis of the information available, historiography has successively developed two rather clear-cut positions on this matter. In the early stages of research, historians and archaeologists viewed the buildings under study as a monumental and official expression of the Aghlabid state’s support for the devotees. Famous formulae, such as that of Ibn Khaldûn mentioning the 10,000 (!) fortresses built by the emir Abû Ibrâhîm Ahmad (reg. 856–863) (Talbi 1966, p. 251), have served as justification for this interpretation. As a result of a historiographical pendulum swing, current research has developed an inverted position that can be considered rather paradoxical: while recognising the role of the central government in the construction of these buildings, historians are almost unanimous in stressing the importance and number of the foundations initiated by private individuals (Djelloul 1999, p. 189; El Bahi 2004, pp. 332, 595, 633; Hassen 2001, p. 153; Amri 2011, p. 339).19

However, the texts and inscriptions that are supposed to support this analysis have been known for a long time; they are very few in number and offer only a limited margin of interpretation. In any case, the available information does not seem to reflect a situation like that which can be observed in the marginal regions of al-Andalus or in the western Maghreb, where the foundation of ribât sites actually reflects the initiative of agents from outside of official circles (Picard and Borrut 2003, p. 56). On the Ifrîqiyan coast, the buildings from the Aghlabid period whose construction circumstances are known are mostly the result of state initiative: the “Qasr al-Ribat” in Sousse was given a minaret by Masrûr, mawlâ (freed slave at the prince’s service) of Ziyâdat Allâh; the Qasr al-Tûb was built by Khalaf, mawlâ of Ziyâdat Allâh I; the emir Abû Ibrâhîm Ahmad had a cistern added to the “Qasr al-kabir” of Monastir; in this same locality, the construction of the Qasr Duwayd is credited to the aforementioned Masrûr on behalf of a member of the reigning family, Duwayd b. Ibrâhîm b. al-Aghlab. In this narrow corpus at our disposal, only three fortresses can thus be attributed to a nongovernmental agent: Qasr Ibn al-Ja’d and Qasr Sahl, built by the eponymous sponsors, and Qasr Ziyâd, built by ‘Abd al-Rahîm al-Rab’î. It is therefore quite easy to see, despite the current consensus among historians on the issue, the importance that the sources attribute to governmental initiative in the founding process of buildings intended to house murâbitûn.

This contradiction in the current historiographical position is perhaps only apparent, and it stems above all from the ambiguous status—private or public—that is given to the euergetistic practices of the persons mentioned in the texts or inscriptions. Current Tunisian historiography tends to consider that the close agents of the prince who are known to us from texts or inscriptions acted privately for the occasion. But, apart from the fact that their position at court places them at the very heart of the government, these same persons or others like them largely contributed to official patronage: these achievements are seen without any ambiguity as public buildings (city walls, mosques, etc.). It is therefore difficult to see why only the ribât sites for which these officials were responsible for supervising the building work and the financing would then appear as private foundations. In the present state of knowledge, the hypothesis of a trend towards the private foundation of ribât sites must therefore be viewed with great circumspection, at least for the most eminent buildings, whose construction method (“stone and lime, with iron doors”, as quoted by Ibn al-Athîr and Ibn Khaldûn, cf. Talbi 1966, p. 251), as well as the overall layout and the fittings, seems to require experienced builders, specific technical facilities and financial resources, all falling within the official sphere (Van Staëvel 2022). Significantly, even in the case of the construction of the Qasr Ziyâd, which is quite obvious, its founder ‘Abd al-Rahîm al-Rab‘î, who was a member of the finest Arab aristocracy, could initiate the construction only after receiving an official decree (sijill) from the emir Ziyâdat Allâh I (Mâlikî, t. I p. 422). It is outside the corpus of the great fortresses where doubt persists. In the case of the more modest buildings (watchtowers, coastal outposts), there is no documentation about the circumstances of their foundation, not to mention the fact that they have never been subject to precise archaeological studies. The heterogeneous nature of these buildings, and consequently the difficulty of attributing them to the Aghlabid period and ribât practices, appears more probably to be a symptom of less centralised construction and a plurality of initiatives.

4.2. A Specific Place at the Core of the Monument: The Mosque

Whatever the circumstances of their erection reported in the texts, the morphology of the major monuments clearly pleads for state intervention according to the modes of implementation proper to official architecture. Another argument in favour of a centralised initiative is the similarity of the known and published buildings: despite certain variations, their repetitive plan shows a high degree of standardisation.20 The quadrangular plan with towers in the shape of a semicircle or three-quarters of a circle, the single rectangular entrance projecting from the middle of one of the four sides and the presence inside of four aisles built on one or two levels against the enclosure wall and opening onto a large central courtyard are all features that have long been familiar to historians of Islamic architecture (Figure 1a–c). The genealogy of these buildings has also given rise to numerous comments, all of which point to an obvious oriental inspiration.21 Indeed, logically, it is the architecture of the 8th-century Near Eastern qusûr that must have served as a source of inspiration and model for the residence of the Ifrîqiyan volunteers of the faith (Lézine 1956, pp. 24–25; Martínez Salvador 1998, p. 251). The architectural formula adopted here in North Africa, however, lends itself to specific arrangements with respect to the oriental paradigm (Van Staëvel 2023, pp. 303–10). The defensive capacity of these establishments was thus greatly increased, incorporating sophisticated devices (slits for throwing missiles or stones, the portcullis) that were typical of early Abbasid architecture (for example, in Ukhaydir or ‘Atshân in Iraq). Similarly, the residential settings are different; while the Umayyad settlements in the steppe are designed with living units (the famous “Syrian bayt-s”) to accommodate large family groups, ribât sites multiply individual cells with no communication between them in order to meet the needs of a large number of male occupants wishing to enjoy their own individual living space.

Finally, and above all, there is one element that distinguishes the architecture of ribât sites in Ifrîqiya, namely, the presence of at least one prayer hall in the most emblematic monuments. This characteristic appears so obvious that, paradoxically, it has not received the full extent of attention in the scholarship, even though it contributes to the specificity of these buildings, which it marks as a particular architectural type.22 This point therefore deserves to be stressed: in its monumental form, at least, the architecture of ribât sites in Ifrîqiya is defined directly by the obvious and unavoidable presence of a mosque.23 This marks the entire building with an undeniable religious seal and also identifies it as a place of prayer. Despite the deep structural and morphological differences between the sites known today in the medieval Muslim West, this characteristic even appears to be a common denominator in many ribât archaeological sites scattered throughout this vast region. Thus, the formal resemblance between the prayer halls (Figure 2) that could bring together the devotees of Sousse, Guardamar or Igiliz on Fridays, during supererogatory prayers or religious festivals, can be highlighted (Ettahiri et al. 2013, pp. 1125–27).24 The similarities do not stop at architectural forms; their dimensions also offer striking parallels from one place to another.25

In the Tunisian ribât fortresses, the degree of integration of the prayer hall into the building is such that it calls for further investigation of the genesis and origin of this unusual architectural feature. In the Near East, none of the monuments known to date have preserved such a layout, and it is therefore quite impossible at this point to identify a supposed prototype. One of the only known archaeological examples that could fit into the category of a ribât site (though not strictly speaking, given its early—Umayyad—date and the absence of any textual reference in this regard) is the coastal fortress of Ashdod/Azdud in Israel. The building is rectangular in plan and indeed contains a small mosque (Raphael 2014, p. 19). But the location of the prayer hall in the middle of the courtyard differs completely from the integrated solution used in Tunisia.26 H. S. Khalilieh also overlooks the need for a mosque in the fortresses of the Palestinian coast that he mentions (Khalilieh 1999, 2008). As we have seen above, no reliable archaeological information is yet available concerning the Syrian–Anatolian thughûr area: this is a huge gap, which weighs heavily on our vision of the architectural phenomenon as a whole. It is likewise difficult to look for an echo of a hypothetical Near Eastern paradigm in the layout of Central Asian ribât sites; the difficult access to archaeological documentation—mainly in Russian, and unevenly available—leaves the question in a state of complete uncertainty. At the most, we can nevertheless point out that the only monument of Paykent published to date as a ribât site does not present any prayer space of the Tunisian type (plan in de La Vaissière 2008, p. 73; see here Figure 1d).27 The current state of our knowledge therefore leaves us with only two alternatives: either early architectural practice, particularly in a frontier context such as Shadod, could have led Near Eastern master builders to conceive a better integration of the prayer hall into the overall plan of the building, and in that case the prototype would have been transferred to Ifrîqiya at the time of the Muhallabids; or, the model of the eastern qasr, imported into Ifrîqiya in its standard form (without a large prayer hall), would have been adapted to meet the specific needs of the local devotees. The second option—by far preferable in the light of the archaeological documentation—could in this case explain the trial and error observed at Monastir in setting up the prayer hall on the first floor of the original building (Lézine 1956, pp. 37–38).28

The need for a common place of prayer obviously depends on the religious practice of the time. Hadîth-s with an essentially devotional content indeed point to an insistence on the performance of prayer (Amri 2011, p. 337). However, it is also possible to put forward another reason, connected to the role that these ribât sites could have played in the religious life of the areas in which they were established. Most studies of these buildings seem to adhere to the idea that their prayer halls were intended solely for their inhabitants. The matter seems all the more clear-cut as these facilities are enclosed within the buildings and can be accessed only from the inside. However, some features of the architecture of the ribât sites in Sousse and Monastir, which have not been discussed much in historiography until now, deserve a closer look. In his meticulous archaeological analysis of the Sousse “Qasr al-Ribat”, Lézine has brought to light the fact that, sometime after the construction of the building, an opening was made in the qibla wall of the prayer hall, giving access to the outside; fitted with a double leaf door, this opening led to a staircase, probably made of wood. This remodelling implies both the end of the military use of the monument and the possibility of direct access to the oratory from the outside.29 How can such a reorganisation be explained?

Historians agree that ancient Hadrumetum seems to have survived in the form of a modest agglomeration (qarya) during the century following the Arab conquest (Djelloul 1999, p. 74). The first sign of a possible repopulation would be precisely the building of the ribât site, under the governorship of Yazîb b. Hâtim (771–788) (Lézine 1971, pp. 85–86). But the urban revival in Sousse really began in 821, with the construction of the arsenal by the Aghlabid emir Ziyâdat Allâh I;30 from that point onwards, the population would only increase (Djelloul 1999, p. 75). The new importance of Sousse was confirmed in 827, when the harbour was chosen as the point of departure for the expeditionary force that was launched against Byzantine Sicily. However, despite this obvious urban growth, it was apparently not until 851 that a congregational mosque was built in the rapidly expanding agglomeration. How was collective prayer performed before then?31 Lézine hypothesises that the “Qasr al-Ribat” may have once served as a mosque for the city’s inhabitants; but he assumes that the access door from the street to the prayer hall could have been opened only after the fortification of the city in 859 had definitively deprived the building of its military function (Lézine 1956, p. 22).32 While following this main idea, we can nevertheless propose an alternative scenario, by moving the building’s functional reform to the beginning of the 820s: the agglomeration, henceforth conceived as the harbour of Kairouan, was then experiencing an influx of Muslim settlers who needed to be offered an adequate place for prayer, and it is the “Qasr al-Ribat” that could have provided it. At the risk that the proposal may seem a little provocative, we might even wonder whether the high tower erected in 821 on the order of the Aghlabid emir should not be seen as a minaret rather than a military watch-post, as it is unanimously interpreted. The striking topographical contiguity between the Great Mosque of Sousse and the “Qasr al-Ribat” would in this case be indicative of the transfer of the function of congregational mosque from one building to the other in the middle of the 9th century.

At Monastir, in the “Qasr al-Kabir”, it has not been possible to bring to light a similar access device due to the lack of excavations. However, Lézine observed the existence of a mihrab on the terrace constituting the roof of the primitive prayer hall: as a result, he believes in the coexistence of two mosques built one above the other (Lézine 1956, p. 38), which probably exceeds the capacity of the original building. He also documented the expansion of the main ribât site during the 9th century, which led to a further increase in the number of prayer spaces in the southeast wing (Lézine 1956, pp. 38–40). Obviously, the presence of these multiple oratories may correspond above all to the growing number of people who joined the devout community, either seasonally or permanently, during the Aghlabid period. However, without overstating this point, one may also wonder whether, during this expansion phase, both the small prayer room located on the ground floor and the one that extends over the entire length of the first floor could also have been accessible to worshippers passing through the inner courtyard. As far as we know, the gateway to this Aghlabid addition dates from the 11th century, as attested by an inscription. However, it would not be illogical to restore an earlier entrance to this same location, corresponding to the original architectural layout. In this case—which only archaeology can corroborate or invalidate—the situation would be more or less the same as in Sousse, with the ribât site playing for a time the role of a congregational mosque for the local population.33

The hypothesis we have just developed raises the more general question of the establishment of the first congregational places of worship in the coastal region of Ifrîqiya, which, according to contemporary historiography, had experienced a strong demographic deflation following the Arab conquest. Unfortunately, we have very little information on this matter. In some parts of the Sahelian coast, it is possible that the ribât sites were the only mosques where the small communities of Muslim settlers could gather: more than a place dedicated to the activities of the ascetic community, in this case the building would be an essential part of the Islamisation process, involving the Muslim populations of the coastal fringe as a whole. Following this hypothesis, we can also better understand the important role played by the state in the foundation of these buildings. As a matter of fact, the construction of Friday mosques is a state prerogative, established by Muslim law. Only the ruler or his representative can legally found a new congregational mosque. He can also authorise a private individual to carry out such a construction: this seems to be the case of the devout ‘Abd al-Rahîm al-Rab‘î, who, as we have seen, requested the authorisation of the Aghlabid emir before starting the construction of the Qasr Ziyâd in 827. One could perhaps see a reminiscence of this role played by the prayer halls of the ribât sites of the first centuries in the recommendation made by the mufti al-Mâzarî (Mahdiya, d. 1141), intended to settle various disputes between individuals, to choose the main mosque of Monastir’s “Qasr al-Kabir” as the place for the solemn performance of the judicial oath (Lagardère 1995, pp. 315–16, 396–97).

5. Conclusions

While noting the remarkable advances made by historiography in recent decades in the construction of a thoroughly renewed image of ribât activities in early Islamic Ifrîqiya, the present contribution has attempted to highlight some persistent limits of analysis or some aspects of the phenomenon that still appear to be underestimated or neglected.

The first, of a general nature, concerns the need to decompartmentalise our perception of the rise of the institution of ribât in Ifrîqiya, in order to inscribe it, by means of a finer contextualisation, within the historical framework of an evolution of the frontier policy initiated by the Abbasid caliphate in the Syrian–Anatolian zone. If the centralising effects of this imperial policy have already been clearly highlighted for Ifrîqiya, the same cannot be said of the parallel but autonomous constitution of a society of warrior ascetics. In Tunisia, this particular social and religious milieu is still treated as a spontaneous generation, specific to this region: on the contrary, the present article argues that it should be reconnected to a wider process of Islamisation. Is this milieu as autonomous as that which developed in the confines of northern Syria in relation to the Abbasid caliphate, or does it maintain more articulated relations with the local power, that of the Aghlabid emirs? Without giving too clear-cut an answer, this article shows that, from the point of view of both the land structures from which the murâbitûn benefited and from that of the sponsorship of the buildings in which they resided, it is the second option that seems the more likely. Without going so far as to speak of a de facto alliance, it nevertheless seems that the two protagonists were able to take advantage of each other to the detriment of a third actor, who has been left out of the picture: the regular army, the jund, which in the Sahel seems to have been pushed aside to a large extent. The rulers sought to win the favour of the men of religion, and the latter undoubtedly made a decisive contribution to the repopulation and development of the coastline, as Tunisian historiography has well noted. The proximity between Kairouan, a political capital and a religious centre of primary importance, and its coastal hinterland obviously played a major role in this particular territorial pattern.

The second point raised by the present contribution concerns the undervalued, but nonetheless utterly central, role played by the prayer hall in the process of founding ribât sites. Surprisingly, although it is eager to discuss the origins of these very characteristic monuments that have been preserved in Tunisia, archaeology has not emphasised the singularity of this architectural device. On the contrary, the present article shows that, at the risk of truism, it is important to put a simple but essential fact back at the very heart of the issue of the material setting of ribât in Ifrîqiya (as elsewhere): above all, the ribât site is a mosque, possibly the only mosque in the area.