Ribāṭ in the Furthermost Coasts of Early Al-Andalus

Abstract

In recent decades, the concept of ribāṭ and its practice have been the subject of intense debate. Recent summary papers on the eastern Mediterranean, Ifrīqiya, al-Maghrib al-Aqṣà, and al-Andalus, among others, have made it possible to compare different realities that express strong links with their local historical contexts. In this paper, we present the results of a new study that analyses the specific case of Northern Sharq al-Andalus, where the practice and institutionalisation of ribāṭ took shape from the early 9th century. There are three elements that lead us to corroborate this hypothesis: the documented presence of numerous individuals and groups voluntarily involved in the active and passive defence of the furthermost frontier of al-Andalus; the confirmation of a construction programme with homogeneous characteristics aimed at building fortified enclosures along the coast, and, lastly, the founding, in the early 9th century, of the Ribāṭ Kashkī centre at the mouth of the Ebro, a building were these practices would become centralised. In conclusion, we propose a much more complex scenario than that proposed previously, which enables us to characterise local forms of armed spirituality and sacralisation of the land that globally enriches the historical reading of ribāṭ.

1. Introduction

In recent decades, the concept of ribāṭ, its evolution over time and the materialisation of its practice in various forms and expressions has become one of the most dynamic areas of historiographical discussion on the early Islamic civilization. As a result, a number of recent summary papers have focused on the in-depth analysis of the genesis and development of this complex reality, especially in the Mediterranean area (Cressier 2019; Morriss 2019; Albarrán and Daza 2019; Van Staëvel 2023). They have taken advantage of the increase in regional studies that have been instigated in various Mediterranean regions over the last two decades, which has led to a growth in local knowledge aimed at heightening the awareness of a global audience about the issues surrounding the concept (Hassen 2001; Cobb 2002; Picard and Borrut 2003; Khalilieh 2008; Eger 2008; Donner 2011; Amri 2011; Chabbi and Rabbat 2012; Mikati 2013; Fenwick 2018; El Bahi 2018).

In general terms, the studies highlight the close link between the concepts of jihād and ribāṭ in the early stages of the spread of Islam. The former emerged primarily in the 8th century onwards and was conceptualised, according to Ibn al-Mubārak, as a spiritual ascetic activity that sometimes entailed taking up arms, which combines striving in the path of God with defending the Islamic faith (Sahner 2017, pp. 149–83; Albarrán and Daza 2019, p. 61). Both concepts are models of praxis and behaviour that link religion and its armed defence within an ideological framework that had already taken shape in Late Antiquity, not only in the Islamic but also in the Christian sphere, and which Islam incorporated and adapted (Sizgorich 2009, pp. 180–86). Of all the concepts surrounding the spread of Islam throughout the 7th and 8th centuries, perhaps the practice of jihād is one of the most important, first, in the push for and later in the active defence of the Dār al-Islām frontier.

The etymological analysis of ribāṭ in its abstract suggests that it is based on the concept of preparing for war, increasing strength, protecting oneself and being ready for combat (Donner 2011, p. 89; Chabbi and Rabbat 2012). It is, therefore, an idea that is strongly linked to, usually passive, defence and the surveillance of frontiers, especially from the 8th century onwards. It developed in parallel with the communal concept of jihād as opposed to earlier, more individualistic understandings (Donner 2011, p. 85). This convention, then, originally alludes to a set of practices that played a key role in the safeguarding and sacralisation of the coast via the mobilisation of volunteers to the frontier, the spread of Islam via its permanent presence in the furthermost corners of the Islamic domains and, lastly, the exegesis of the action of the state. In essence, ribāṭ and its practitioners materialised as a way to culturally construct a sacred frontier space, enabling Islam to take root via different social and political realities (Munt 2014, pp. 8–15). It was only at a later stage that the word would become a synonym for a place where ribāṭ was performed (Morriss 2019, pp. 37–38).

The perfect space for the embodiment of both concepts was the frontier with Dār al-Ḥarb, an area of physical and ideological confrontation that required the permanent engagement of not just Muslim rulers but also believers. Additionally, the Islamic state of al-Andalus fully responded to the context as an island surrounded by seas and hostile lands. A place where the need to protect the conquests of Islam was ever-pressing. Home of jihād and ribāṭ, as attested by the Arab authors themselves (García Sanjuán 2016, p. 45), the practice of both in al-Andalus is linked to an active defence of Muslim frontiers while practising vigilant ascetic spirituality (de Epalza 1993; Franco 2010). The al-Andalus treatises that deal with this subject, especially the work by Ibn Abī Zamanīn (d. 1008), highlight the importance of the murābiṭūn or practitioners of ribāṭ, which he describes as ascetic defenders, in protecting the territories gained during the Islamic expansion (Albarrán and Daza 2019, pp. 62–64). Ribāṭ is, ultimately, a fundamental component in the configuration of the territories’ frontiers, especially in those sectors in which land and maritime boundaries were united in a single conflictive space.

Despite the convincing written testimonies that help us to define the two concepts, the material evidence expresses a multifaceted and much more diverse reality than the one put forward so far. The practice of ribāṭ entailed the development of a more varied set of activities than those that could be expected from a group of locations whose sole function was linked to military practice. It is precisely this reality, a reflection of commercial, economic, social and religious interests, among others, that we aim to highlight in this paper. A scenario that would see the far-flung shores of the Sharq al-Andalus provide the ideal conditions for early settlement and contemporary experiences similar to that of Ifrīqiya. By recognising their influence in the process of Islamisation on the margins of this new world in expansion, we can extract the keys to understanding the important role these practices played in the genesis of a new Islamic society in the western Mediterranean.

The Study of Ribāṭ in the Mediterranean

Approaches to the study of the phenomenon of ribāṭ along the Mediterranean coast over recent decades have typically tended to have a strong military bias. This is due, in large part, to the extensive network of various kinds of coastal fortifications developed and/or improved as the Islamic conquest spread along the eastern Mediterranean coastline. Both Arab authors and later historiographers have provided a homogeneous perception of these fortifications, linking them to a greater or lesser extent to this perception (Morriss 2019, p. 38). However, the network was actually the result of the maritime policy of the early caliphs after the conquest of Syria and Egypt, which was aimed at building or repairing fortresses on the coast as opposed to the more sheltered position of the main administrative centres. This encouraged populations and military contingents to settle in coastal regions and provided an early warning surveillance system (Lirola 1993, pp. 58–59). In this context, the simplification of the varied and changing notions of ribāṭ to a single concept of military architecture, or the unification, under this concept, of different forms of imposition on the territory, has for decades weakened the reading of a multifaceted and complex reality.

It is not only in this respect that the study of ribāṭ as one of the earliest expressions of Islam’s armed and spiritual defence is severely limited; the over-reliance on written sources and their apocalyptic accounts of the emptiness of frontier territories is also a limiting element, as is the anachronistic use of certain descriptions produced centuries after the advent of the concept (Morriss 2019, pp. 35–36). Archaeological studies are equally unhelpful as most of the examples documented in written sources have been destroyed, absorbed by later buildings or, in some cases, are still awaiting excavation. Despite this and based on a few archaeological excavations and idealised construction plans, the search for an architectural model capable of explaining all the buildings linked to mentions of rubūṭ has been a major, unrelenting research objective for years (Marçais 1925; Oliver 1928; Lézine 1966).

In contrast, more recent studies characterise the earliest expressions linked to ribāṭ, taking into account the polyhedral reality which surrounds the concept. The studies include the analysis of places linked to ribāṭ practices in the eastern Mediterranean, Ifrīqiya and the al-Maghrib al-Aqṣà, and reveal diverse and complementary functions (Khalilieh 1999, 2008; Hassen 2001; Picard and Borrut 2003, p. 45; Fenwick 2018, pp. 216–17; Morriss 2019, pp. 44–51; Cressier 2019, pp. 117–20).

In general terms, the preponderance of the defensive element during the early phases is notable, even though it coexisted with other functions of an economic, social or religious nature. The military element gradually diminished in importance to give way to new forms of spirituality and the relationship between locations and their surroundings. In short, cities, fortifications and watchtowers, among other places, were considered rubūṭ when they and their immediate surroundings were immersed in the dynamics of the concept’s practices. Therefore, initially, there was not a single type of building known as a ribāṭ, but rather various constructions integrated within the scope of the concept and its variable objectives as a focus for the defence and dissemination of the Muslim faith and model of Islamic society.

From interaction with local communities wherever ribāṭ was established, a process of hybridisation emerged between the different characteristics from each scenario, which gave rise to analogous realities rich in nuances and details. Many authors highlight other functions in the development of locations associated with ribāṭ, such as their capacity to absorb and integrate the objectives of earlier buildings located in any given territory. They were, therefore, the manifestation of the state’s power and propaganda narrative, in the same way as the eastern castra, from which they also adopted other functions, such as monitoring communication networks, maritime surveillance and managing adjacent territories (Hassen 2001, p. 154; Picard and Borrut 2003, p. 39; Genequand 2006, p. 25; Eger 2008; Fenwick 2018, pp. 216–17; El Bahi 2018, p. 323; Bosanquet 2023, pp. 6–7).

These facilities were also stopping places and logistics hubs on major transport networks, including maritime routes, which enabled them to control the flow of goods and tax payments (Khalilieh 1999, pp. 217–19; Eger 2012, p. 445). Closely related to this, they also played a key role in the development of peasant coastal settlements; some of them linked to ribāṭ became actual cities (Hassen 2001, p. 153; El Bahi 2018). Moreover, eastern locations played a key role as places for the exchange of prisoners, a function that would essentially be performed beyond the reach of state intervention (Friedman 2002, pp. 34–35; Morriss 2019, pp. 50–51). Far from a preset, static perspective, these sites absorbed a diversified amalgamation of objectives, which developed and disappeared according to their immediate context.

Although less research has been performed, the scenario on the coast of al-Maghrib al-Aqṣà and the Mediterranean coast of al-Andalus is equally complex and in desperate need of new approaches. This has been highlighted by Cressier (2019, pp. 120–21) in the case of Africa’s Atlantic coast and Albarrán and Daza (2019, pp. 91–95) in the case of al-Andalus.

Dating from the 9th to the 10th century, the reasons behind the founding of centres linked to ribāṭ on the western coasts of the Maghreb include Norman invasions, the presence of heterodox communities such as the Barghawāṭa and the proselytise practices of certain important figures. The spiritual and commercial role of these centres can be seen through the founding of places of worship and periodic fairs or markets, while their military or defensive function is practically imperceptible (Cressier 2019, pp. 120–21).

In the case of al-Andalus, numerous analyses have been performed, especially in the field of Arab Studies (Oliver 1928; de Epalza 1993; García Sanjuán 2000; Marín 2004; Martínez Salvador 2004; Franco 2010, among others). As regards written sources on the subject, there has been much interest in Pavón’s quandary, i.e., the search for a common model pertaining to the concept, which he concludes does not exist in the case of al-Andalus (Pavón 1995, pp. 128–29). Despite this, toponymic, epigraphic and written evidence has enabled various researchers to propose a relationship between some buildings, usually fortified or monumental constructions, and this type of reality, as summarised by Albarrán and Daza (2019, pp. 71–79). For the cases where it has been argued, with a certain degree of certainty, that they were places of ribāṭ, a series of criteria have been established which, in combination, enables us to propose not one but several models for these foundations. Possible lines of evidence include the existence of defensive structures, a place of worship, homogeneous living areas around an open space, written sources that mention the practice of ribāṭ, verifiable historic toponymy and, lastly, their location in ‘conflictive’ areas and access routes for the state’s incursions.

The confirmation of a fluid model that links buildings with the practice of ribāṭ is an important step forward and breaks with simplistic readings. Even so, there is still a pressing need for new case studies to be performed in order to test this theory. One of the most relevant examples in this regard is that performed in Guardamar (Alicante, Spain), which, even without exhausting its archaeological potential, has provided clarity on a certain specific type of construction. Built-in the late 9th century by the baḥriyyūn or seafaring communities from Pechina (Almería, Spain) as a sporadic trading post, the building was transformed during the Caliphate period into a cultural and religious centre designed to facilitate prayer and spiritual retreat, with cells and oratories around a mosque and spaces located outside the sacred area to accommodate pilgrims and travellers (Azuar 2005, 2019). Overall, this validates the proposal put forward to identify rubūṭ by verifying several of the established criteria. In this case, a testimony in written sources, the existence of several buildings for worship and a specific layout for residential areas. Their particularities, as we defend in this paper, respond to the set of local social, economic and religious conditioning factors on which the concept is founded.

At this point, it is worth questioning whether, in the area under study, it is possible to establish not just one specific type of building intended for the practice of ribāṭ but, as we suspect, a complex network of buildings linked to the concept. In this regard, in the case of 9th-century Northern Sharq al-Andalus, around the domains of Tortosa, the preliminary results are very promising. Therefore, it is important to assess the capacity of written and archaeological sources to facilitate a precise definition of this framework and its relationship with the spread, consolidation and defence of Islam in its furthermost frontier.

2. On the Dissemination of Ribāṭ in Northern Sharq Al-Andalus

During the spring of 711 CE, Muslim armies arrived in the Iberian Peninsula, taking advantage of a climate of deep animosity between different sectors of the Visigothic aristocracy after the death of King Wittiza (Manzano 2006, pp. 32–33). The troops comprised a cadre of mostly Arab officers but also a large number of members from Berber tribes: Maṣmūda, Zanāta, Hawwāra, Madyūna, Maṭghara, Jarāwa and Kutāma, among others, commanded by Ṭāriq b. Zīyād (Ibn ʿIdhārī 1948–1951, vol. II, pp. 9, 126–28; Ṭāha [1989] 2016, p. 87). The campaigns against coastal Tarraconensis took place from the very beginning of the invasion, if written sources, which describe various attacks beyond Zaragoza and in Tarragona itself, are to be believed (Al-Maqqarī 1988, vol. I, pp. 273–74; Al-Rāzī 1975, p. 41; Al-Ḥimyarī 1984, pp. 391–92). The return of the campaign leaders to Damascus, as well as the various fronts opened during the early years of the conquest, did not facilitate the consolidation of the eastern territories of Tarraconensis and Narbonensis until Governor al-Samḥ was appointed to the office from 719 to 721 CE (Chalmeta 1994, pp. 259–68; Manzano 2006, pp. 43–44).

Coastal Tarraconensis continued to maintain a certain character of its own throughout the 8th century, as manifested by the city’s control of extensive domains that stretched from Narbonne in the north to Tortosa in the south (Negre et al. forthcoming). The Umayyad general Abū ʿUthmān Ubayd Allāh b. ʿUthmān, known as ṣāḥib al-arḍ al-kabīra (lord of the great land), established himself in the former Roman provincial capital during the middle decades of the century from where he collected taxes from the main coastal cities on the furthermost frontier of al-Andalus, especially Barcelona and Narbonne (Al-ʿUdhrī 1965, pp. 27–29; Ibn ʿIdhārī 1948–1951, vol. II, p. 65; Al-Ḥimyarī 1984, p. 389; Lévi-Provençal 1953, p. 77; Sénac 2015, pp. 44–45). However, from the second half of the 8th century, the city of Tortosa itself also began to gain importance in the territory, as reflected in the appointment of its governors. This is the case of ʿAbd al-Raḥmān b. ʿUqba and Tammām b. ʿAlqama al-Thaqafī, who exercised their authority over several cities at the same time, and ʿUbaydūn b. al-Ghamr, who was a client of the plenipotentiary ʿAmrūs b. Yūsuf and later became governor of Zaragoza, as would his son sometime later (Ibn al-Qūṭiya 2009, p. 72; Ibn al-Abbār 1963–1964, vol. I, p. 143; Ibn Ḥayyān 2003, pp. 102, 132, 179–80; Ibn ʿIdhārī 1948–1951, vol. II, p. 74).

It was precisely the figure of ʿUbaydūn b. al-Ghamr was key to understanding the development of the frontier lands of Tarragona and Tortosa at the turn of the century. The Islamic successes in Frankish territory in the early 8th century were quickly responded to by the Carolingians, who managed to recover the territory lost in a matter of decades. By the end of the century, several important towns in coastal Tarragona had surrendered to the Christian drive, including Girona shortly after 790 CE and Barcelona in 801 CE (Astronomus 1995, pp. 312–20; Ibn Ḥayyān 2003, pp. 116–17). This success added to the situation generated by Charlemagne that was largely beneficial for the new empire he had built in the heart of Europe. He was not prepared to slow down his war machine until he had driven the Muslims beyond the Ebro Valley. Moreover, the campaign was supported by Frankish aristocracies, who were eager to stop the attacks that the al-Andalus ships were launching against the coast and the ships of south-eastern Gaul (Nigellus 1932, pp. 12–15; Sénac 2002, p. 71; Ballestín 2014, p. 69).

Between 804 and 809 CE, the same argument led the Carolingians to organise a frustrated campaign on the Ebro, in this case with the aim of conquering Tortosa, the city from where the attacks now started, but which was able to resist the Frankish raids on three occasions (Annales 1895, p. 127; Astronomus 1995, pp. 320–31; Suñé 2016, pp. 978–82). The governor of the city, ʿUbaydūn b. al-Ghamr played a key role in these campaigns by organising the city’s defences and leading the border and irregular troops to prevent Charlemagne’s army from crossing the Ebro. Everything seems to indicate that, with multiple fronts open and an army with limited resources, defending the borders of the capital of the Ebro rather than the huge megalopolis of Tarragona was a less risky choice for Córdoba to organise (Negre et al. forthcoming). Please see Figure 1 below.

After the failure of the Frankish sieges, control of the area was consolidated under the Umayyad Emirate, while the Carolingians had to retreat as far as Barcelona, ultimately desisting from expanding their dominions as far as the Ebro Valley. This scenario redefined the limits of the frontier between the Christian and al-Andalus territories, establishing the river Llobregat as the area of contact between the two political structures (Al-Maqqarī 1988, vol. I, p. 128). This situation forced not only the emir but also the entire Muslim community of al-Andalus to rethink its strategy for the effective defence of its furthermost frontier, now centralised in the city of Tortosa (Negre and Suñé 2019). This maritime and land border represented the spearhead of Dār al-Islām in the Christian West, but also perhaps the Muslim most exposed sector to possible counter-offensives.

2.1. The Ribāṭ and Jihād Volunteers

Three elements seem to support the hypothesis that there were volunteers dedicated to the practice of ribāṭ on the coasts of Tortosa from the 9th century onwards. First, the documented concentration of large communities of baḥriyyūn, al-Andalus seafaring communities who were prepared for direct action or to support operations linked to the practice of jihād, especially those of a maritime nature, on the coasts of Northern Sharq al-Andalus from the late 8th century. Second, the participation of muṭṭawwiʿa or irregular volunteer troops among the defenders of Tortosa during the last Carolingian attack on the city in 808–809 CE. Lastly, the presence of Tortosa of murābiṭūn, confirmed by the construction of the Ribāṭ Kashkī at the mouth of the Ebro at the beginning of the century, a subject that is discussed later in this paper.

The baḥriyyūn, literally ‘people of the sea’, were communities of sailors from the coasts of ancient Tarraconensis, who, from the late 8th century and throughout the following, were able to organise important naval expeditions (Negre forthcoming). They were seafarers whose ventures are documented as having first set sail from the city of Barcelona, although shortly afterwards, they were forced to move to Tortosa and later spread to the rest of the Sharq al-Andalus coastal areas (Guichard 1983; Lirola 1993; Picard 2007, pp. 9–20; Ballestín 2014; Negre and Suñé 2019).

The most detailed information about these different groups and their actions throughout the 9th century is provided by al-Ḥimyarī, whose works comprise numerous scattered reports whose origins have been partially blurred by the passage of time. Among other details, the Arab author wrote that the baḥriyyūn belonged to the lowest strata of society, lived mostly in the region of Tortosa and were engaged in the maritime transport of people, traded various products—including luxury goods—and plundered the coasts and ships of the infidel (Al-Ḥimyarī 1984, p. 80). These latter activities are included in the group of practices that constitute the concept of jihād, which is why some sources identify some of these crews as mujāhidūn (Ibn Ḥawqal 1992, pp. 174–75; Liutprand of Cremona 1839, pp. 275–76), a term that can be considered analogous to murābiṭūn in a different context of armed conflict.

It is therefore worth noting the apparent correlation between the settlement of communities dedicated to sailing and maritime trade and the founding of rubūṭ at various points along the coasts of the Maghreb and the Iberian Peninsula (Cressier 2004, pp. 207–9; 2019, pp. 118–20; Azuar 2019, pp. 132–34; Bosanquet 2023, p. 10–11). The baḥriyyūn were nothing more than the material expression, in this case, linked to al-Andalus, of Mediterranean seafaring traditions that were progressively integrated into the Islamic social fabric via various transculturation mechanisms. Traditions which, throughout antiquity, had allowed permanent trade routes to flourish between the coasts of the Iberian Peninsula and those of the Maghreb and which were still active throughout the 8th century, as evidenced by several recent archaeological excavations performed in port areas (Rodríguez et al. 2020; Negre et al. forthcoming; Lasheras and Rodríguez forthcoming).

Thus, these groups of sailors would have set sail from the main ports of ancient Tarraconensis, especially Tarragona, Barcelona and Tortosa. However, the expeditions instigated by these people devoted to the sea and sailing did not appear out of nowhere but formed part of a well-established guild on the coast of the Iberian Peninsula that had not stopped its activity since ancient times (Ibn al-Qūṭiya 2009, pp. 57, 61; Lirola 1993, p. 83). These crews included not only local sailors but also progressively incorporated numerous individuals belonging to newly arrived groups. All in all, a collective with good knowledge of the al-Andalus and African coasts would have favoured not only commercial exchange but also the transmission of knowledge, traditions and new forms of spirituality, such as ribāṭ itself (Negre forthcoming).

Among other notable jihād actions performed by these communities from the northern coast of Sharq al-Andalus were the expeditions of Mallorca and Menorca (798 CE), Baleares (799 CE), Corsica (806 CE), Pantelleria (806 CE), Sardinia and Corsica (807 CE), Corsica (809 CE), Sardinia and Corsica (810 CE), Corsica (810 CE), Sardinia (812 CE), Mallorca (813 CE), Civitavecchia and Nice (813 CE) and Alexandria (815 CE) (Al-Ḥimyarī 1984, p. 429; Ibn ʿIdhārī 1948–1951, vol. I, pp. 96–97; Annales 1895, pp. 104–5, 108–9, 122, 124, 128, 130, 133, 137, 139; Ibn Ḥayyān 1973, pp. 2–3; Aguadé 1976, pp. 162–65; Guichard 1983, pp. 60, 62–63). Additionally, although from 815 CE onwards, a significant number of al-Andalus sailors moved to the eastern Mediterranean to pursue their way of life more freely after the Frankish coastal defences were reinforced, their activity in the west did not stop altogether. There are indications that point to small raids against Marseille (838 CE), Arles (842 CE), Rome (846 CE) and Arles for a second time (850 CE) (Aguadé 1976, pp. 162–63, 165; Guichard 1983, n. 72; Lirola 1993, p. 151).

Furthermore, sources highlight two campaigns: that of 829 CE, during which more than three hundred ships set sail from the Tortosa coast and actively collaborated with the Banū l-Aghlab in the conquest of Sicily (Al-Ḥimyarī 1984, p. 429; Ibn ʿIdhārī 1948–1951, vol. I, pp. 96–97; Lirola 1993, pp. 107–9, 138; Ballestín 2014, pp. 71–72) and the campaign of 848 CE, with another group of three hundred ships which, on the initiative of Emir ʿAbd al-Raḥmān b. al-Ḥakam was sent to punish the inhabitants of the islands for breaking the pacts of submission and attacking al-Andalus ships (Ibn Ḥayyān 1973, pp. 2–3; Ibn ʿIdhārī 1948–1951, vol. II, p. 91). Please see Figure 2 below.

Overall, the organisation of such major expeditions had to rely on various factors to ensure their success, such as the mujāhidūn, volunteers willing to embark on or sponsor these major operations, and the baḥriyyūn, without whom the ships could not be manned, as well as the collaboration of the state authorities, via the approval and support of the emir or his provincial representatives (Negre and Suñé 2019, pp. 714–15). These three elements, which were essential for the favourable development of any large-scale campaign, are clearly seen in a later operation, many of whose details have, fortunately, been preserved in Ibn Khaldūn’s description of the conquest of Mallorca in 903 CE. The campaign owes its success to the collaboration between Emir ʿAbd Allāh, despite being embroiled in a deep institutional crisis, and the private initiative of a wealthy and devout Muslim, ʿIṣām al-Khawlānī. Both commissioned numerous ships and recruited volunteers willing to embark, although the leadership of the expedition was left solely in the hands of the latter. On his return, after the success of the operation, ʿAbd Allāh rewarded him by making him governor of the island (Ibn Khaldūn 2000–2001, vol. IV, p. 210).

There is no information on new expeditions beyond the mid-9th century, although the activity performed by these groups must have continued throughout the period. They were probably neglected by courtly sources who were busy communicating the central government’s leading role. Written evidence also indicates that at some point before 875 CE, a group of sailors from Tortosa, Elvira and Tudmir decided to join forces and violate the maritime rules legislated by the Umayyad emirs, to the extent that they attacked the al-Andalus city of Marchena (Al-Ḥimyarī 1984, p. 80). Realising that their return to the shores of Sharq al-Andalus would not be safe, they set course for the North African coast, where they participated in the transformation of the fortress of Tenes into a small medina, which eventually became the main centre of gathering and exchange for the Berber groups in the area.

It was not until a few years later, around 884 CE, that a large group of baḥriyyūn who had settled in Tenes, a good many of whom were originally from Tortosa, returned to the Iberian Peninsula (Ballestín 2014, pp. 72–73; Picard 2007, pp. 432–42). Taking advantage of the general instability of the revolts in the late 9th century, their destination was the region of Pechina, from where they managed to expel the Arab governors and take control of the territory after gaining the emir’s approval. However, their exile and experience in Tenes seem to have changed their priorities, and they saw the crisis of the Emirate as an opportunity to establish themselves as a local power capable of setting the course of their future actions (Negre forthcoming).

The way these communities of sailors developed their operations throughout the 9th century had a profound effect on al-Andalus maritime endeavours and the participation of volunteers in jihād prior to the Emirate organising a navy of its own. However, once the latter appeared on the scene, the baḥriyyūn gradually abandoned their warlike facet in order to definitively strengthen their commercial virtues, which had been present from the very beginning.

In contrast, the muṭṭawwiʿa were irregular troops who were not paid a public salary for their involvement in military campaigns and who, at least in the east, were sent to the frontier in order to participate in jihād military raids (Chalmeta 1976, pp. 414–15; Meouak 1993, pp. 369–70; Aguilar 1997, p. 196; Molénat 2005, p. 555; Viguera 2001, pp. 27–28). Ibn Ḥayyān states that during the Carolingian attacks against Tortosa, they were under the command of the governors of Zaragoza and Tortosa, a circumstance that suggests that they were already at the front before the Cordovan troops came to the city’s aid (Ibn Ḥayyān 2003, pp. 131–32). This, together with the well-known fact that several ulema and Muslim believers who were willing to sacrifice themselves for Islam were among the contingent’s members (Ibn Ḥayyān 1965, p. 226; Ibn ʿIdhārī 1930, p. 4), clearly links the muṭṭawwiʿa with the practice of ribāṭ, which went further than their occasional adherence to jihād offensive campaigns over the border (Suñé 2016, pp. 984–85; Negre and Suñé 2019, pp. 717–19).

In general terms, these groups of volunteers do not seem to have settled in entirely isolated places, nor to have performed early forms of military asceticism in a totally bespoke way, but rather to have collaborated with local authorities and taken advantage of a stable infrastructure that enabled them to combine surveillance, combat and prayer more effectively (Albarrán and Daza 2019, pp. 70–71). In this regard, the fortifications and architectural complexes that we analyse in the following sections might have responded to those needs. From there, they could defend the frontier and recruit reinforcements during jihād campaigns against the territories beyond the al-Andalus frontier. This is what seems to emerge from the analysis of the al-Andalus campaigns against Christian territories throughout the 9th century and during the fītna (Suñé 2017, pp. 128–29), where voluntary contingents appeared on a regular basis (Ibn Ḥayyān 1973, pp. 271–73; 1979, pp. 88, 95, 125; 2003, p. 136; Ibn ʿIdhārī 1948–1951, vol. II, pp. 180, 199). However, the bulk of the army was formed by border troops and people from local towns and fortresses (Suñé 2017, pp. 138–39), which leads us to err on the side of caution when assessing their involvement in offensive actions.

Conversely, not all volunteers were anonymous individuals or faceless groups. At least seven well-known figures were clearly identified in written sources as fuqahā who had been based in Tortosa during the 9th century, three of whom died as a result of fighting against the Christians. They were Ṭāhir b. Ḥazm and Abū Zakarīyāʾ Yaḥyā, blood relatives who lost their lives in a raid against Dār al-Ḥarb, and Abū al-Muṭarrif ʿAbd al-Raḥmān b. Muʿāwiya, who would reach Jannah together with two of his sons fighting against the Christians at the front in Barcelona (Ballestín 1994, pp. 67–73). This gives rise to the idea of a reality on the frontier that was profoundly spiritual, where not only was it righteous to defend the institutions of Islam but also to pay the ultimate sacrifice via jihād. This situation might have seemed commonplace for people with a deep religious vocation who had gone there with the aim of acting as beacons of Islam.

From the above, it can be inferred that there was no shortage of volunteers on the shores of Tortosa who, individually or collectively, went to the furthest frontier of Dār al-Islām to perform these new forms of armed spirituality throughout the Emirate period. Murābiṭūn, mujāhidūn, muṭṭawwiʿa and baḥriyyūn were just some of the faces of the same multifaceted reality: that of thousands of individuals willing to pledge or even sacrifice their lives and worldly goods in order to defend Islam against its enemies. However, other less devout and more materialistic motives should not be ruled out. Large communities flourished or were transformed in the midst of this practice, and although not all of them played an active role in the armed struggle, they performed other much-needed, essential support services. Passive or offensive defence, ribāṭ or jihād in the face of Dār al-Ḥarb, were two accounts of the same narrative built on new individual and collective religious expressions, typical of a militarised and deeply sacralised frontier.

2.2. The First Archaeological Evidence of the Practice of Ribāṭ: A Fortified Geography

The ancient territory of coastal Tarraconensis, which from the beginning of the 9th century came under the sphere of influence of the city of Tortosa, harbours one of the most unique al-Andalus fortified networks in the Iberian Peninsula. Generally speaking, it reflects an occupation dynamic characterised by certain sites located at high altitudes, with clear parallels in the rest of the western Mediterranean during Late Antiquity and the Early Middle Ages (Schneider 2001; Catalán et al. 2014; Gibert 2018). Moreover, these fortifications responded to very diverse social and political realities, which owe their existence to multiple causes. This is why complementary readings have been proposed to explain the founding of these enclaves and their role in history: as centres of local power (Schneider 2001, pp. 436–39; Castellanos and Martín Viso 2005), as road checkpoints (Schneider 2001, pp. 440–42; Bazzana 1992, p. 277) or even as settlements linked to peasant communities that tried to escape the control of major territorial powers (Bazzana 1992, p. 280; Acién 2001, p. 60; Schneider 2001, pp. 439–40), among other explanations. Please see Figure 3 below.

In the specific case of Northern Sharq al-Andalus, we owe a great deal to André Bazzana for illuminating the issue (Bazzana 1992, pp. 273–75). However, any further development of the subject had been relatively neglected until a decade ago. The French archaeologist’s initial proposal focused on identifying the group of fortifications by using a series of common features, such as their construction using drystone masonry, their location on top of topographically important peaks along the coast and communication routes, and the existence of structures or materials that ascribe them to the Emirate period. He also differentiated between enclosures with a higher degree of complexity and others, described as enigmatic in his study, with simpler characteristics (Bazzana 1992, pp. 348–53).

In the first group, he highlighted the cases of Mollet (Sant Joan de Moró, Castellón, Spain) and Marinet (Xodos, Castellón, Spain), for which he suggested different functions: those of a military enclosure linked to road and territorial control (Bazzana 1992, pp. 277–78), and those of a settlement-shelter for peasant populations (Bazzana 1992, p. 280). The French author distinguished two clearly separate spaces in the former, which is located on a high promontory near the coastal road between Valencia and Tortosa: the first, built on top of one of the highest peaks, is a fortified enclosure of around 3000 m² with a complex internal floorplan organised around a central diaphanous space and a redoubt which may have originally been a tower and, the second comprises various living and service structures in a larger area located a few metres away on a lower level (Bazzana 1992, pp. 277–78). In the absence of new and much-needed excavations at the site, especially in the totally unexplored fortified enclosure located on the upper level, the technical specifications, poliorcetic techniques and material records suggest a space clearly used for military purposes, although the detail of its internal spatial organisation is still unknown.

The second example is a high-altitude settlement (1467 m above sea level) located on a rocky outcrop that outlines a triangular-shaped floorplan flanked by narrow cliffs on its southwest and east sides and enclosed by a monumental drystone masonry wall on the north side. Archaeologically, different sectors can be distinguished, which are marked by the topography of the site, with its steep slopes. First, the inner face of the wall has a series of quadrangular living spaces arranged in two parallel lines attached to the perimeter defence. The rest of the platform is covered by a number of simple structural areas, which have no discernible differentiated distribution between them. Interestingly, a larger building structured around a courtyard was identified at the southeast end, while on the central, higher level, the foundations of a thick-walled quadrangular building were documented, which may have originally corresponded to a tower (Bazzana 1992, p. 280).

In the second group, he could only define a series of common characteristics: a medium-altitude location with difficult access, their construction exclusively using drystone masonry, a perimeter wall adapted to the contour lines, the absence of living quarters, and very few pottery finds. He linked these structures to pastoral activities, such as those attributed to the sites of La Garrotxa and Subarra, to which new cases were added in subsequent years (Bazzana 1992, pp. 351–53; 2000, pp. 508–10).

Owing to recent survey work and the revision of museum collections, but also to systematic archaeological excavations, we now have a considerably extended list of sites, which presents a panorama that is clearly richer in detail. To Bazzana’s first group of sites in which living spaces and greater complexity in poliorcetic techniques were identified, we have added the paradigmatic and extensively excavated case of Tossal de la Vila (Negre et al. 2020a, 2020b, 2021, 2022; Pérez-Polo et al. 2023a, 2023b), as well as other cases that have similar characteristics, which are still under study, i.e., El Castellar and La Mola de Genessies. As regards the second group of sites, we could add other examples of large enclosures that have no internal structure, including El Gaidó, El Morico, El Puig de la Mola and El Puig de Gallicant (Bazzana and Guichard 1979; Miret 2011, pp. 224–27; Gibert 2018, p. 159). This is far from being a closed or minimally definitive corpus, as the signs of more examples increase with each extensive intervention in the area.

From the sites where more intensive excavation has been performed, it is worth highlighting the case of Tossal de la Vila (La Serra d’en Galceran, Castellón, Spain). This site is located on one of the most prominent peaks (954 m) in the south of the Serra d’En Galceran, a mountain range that delimits two of the main communication routes in the central area of the province of Castelló, Spain. It spans a wide rocky platform of around 5000 m², enclosed by a perimeter wall located on the north, west and south sides of the settlement, while the eastern side of the settlement was naturally defended by escarpments. Based on the absolute dates obtained from various biological samples, the archaeological record recovered, and the various construction phases that have been documented at the site, we can confidently suggest that the enclosure was founded between the late 8th and the early 9th centuries. It would have been abandoned in a planned and calculated way during the first half of the 10th century, most probably at the end of the first quarter of the 10th century. Please see

Table 1. Periods of al-Andalus occupation at Tossal de la Vila.

Period	Phase	Chronology	Dating Criteria
Emirate period	Construction	Late 8th c.–Early 9th c. CE	Ceramics and stratigraphic
Emirate period	Use	Early 9th c.–Early 10th c. CE	TV-1; TV-2; TV-3
Fitna-Caliphate period	Abandonment	Early 10th c. CE	TV-4; TV-5

and

Table 2. Absolute al-Andalus dating of Tossal de la Vila.

ID	Laboratory Code	Sample Type	Conventional Radiocarbon Age (BP)	Calibrated Age (1σ 68.2%)	Calibrated Age (2σ 95.4%)	Calibration
TV-1	Beta-588420	Charred material	1220 ± 30	784–835 cal CE 844–877 cal CE 774–778 cal CE	770–888 cal CE 686–742 cal CE	IntCal20
TV-2	Beta-503127	Charred material	1200 ± 30	782–192 cal CE 801–811 cal CE 819–881 cal CE	771–894 cal CE 706–736 cal CE 929–945 cal CE	IntCal20
TV-3	Beta-639687	Bone collagen	1160 ± 30	920–955 cal CE 870–898 cal CE 828–861 cal CE 776–787 cal CE	820–978 cal CE 772–790 cal CE 804–810 cal CE	IntCal20
TV-4	Beta-588421	Bone collagen	1130 ± 30	915–976 cal CE 889–902 cal CE	875–994 cal CE 830–851 cal CE 776–786 cal CE	IntCal20
TV-5	Beta- 419185	Charred material	1050 ± 30	987–1026 cal CE	895–925 cal CE 949–1035 cal CE	IntCal20

below.

Initially, a formidable 2–3 metre-wide wall was built on the ruins of protohistoric structures along the most easily accessible sides, which was interspersed with towers and bastions. Regarding the structure of the enclosure’s interior space, we only have evidence of buildings in the spaces under the protection of the wall, the latter acting as the interior facing of a series of simple structures built in a battery facing a central open space (Negre et al. 2020b, pp. 157–60). Two structures are the only free-standing buildings documented at the site. One is a rectangular living space, identical to the simple structures attached to the wall but built using clearly better construction methods, and the other is a large building with a surface area of 60 m² that has been identified as a mosque (Negre et al. 2022). This place of worship is, to date, one of the earliest Islamic rural temples documented in the Iberian Peninsula. Its exceptional nature, including two maḥārīb perhaps intended to segregate the room into two separated prayer spaces by means of a movable unit and a qibla that is perfectly aligned with the sunrise at the summer solstice, makes it a highly relevant case study.

If we consider the architectural evidence, the analysis of the material contexts that were recovered and the enclosure’s strategic location at the intersection of the roads leading from the city of Valencia to Tortosa and Zaragoza, it leads us to the conclusion that the hillfort was an enclave with a clear military vocation. This is also indicated by a wide and varied collection of weapons and pottery finds, some of them made in centres far from the enclave, such as Tudmīr, Córdoba and the Maghreb (Negre et al. 2020b, pp. 205–9).

There is also no evidence of agrarian practices developed by the inhabitants of the site, while the intensive exploitation of the biotic resources available in situ is instead well-attested. In this regard, the economic activities documented in its interior were the minimum needed for the subsistence of the group, linked practically and exclusively to providing firewood for the living spaces and maintaining a small number of sheep, goats and chickens to provide a constant supply of food for the garrison. The archaeozoological evidence also seems to highlight that careful choices were made via the intensive exploitation of animals and their by-products, such as milk and eggs, as sources of animal protein. This seems to indicate that agrarian activities in the settlement were reduced to a minimum, suggesting that the economic strategies favoured the simplification of the work processes linked to subsistence. This enabled efforts to focus on the military functions that were expected of a fortification of this type (Pérez-Polo et al. 2023b, pp. 21–23).

The abandonment of the site, detected using sequence stratigraphy, as well as the constant alterations to its living and storage spaces, indicate a use focused on settlings of unknown and intermittent duration by groups of occupants who were not excessive in number, perhaps around 50 individuals (Negre et al. 2020b, p. 161).

If we accept the above, other highly important archaeological evidence was also detected during the excavation of the site that could reveal some nuances about the idiosyncrasies of the contingent installed there. All the excavated elements in the enclosure indicate an initial egalitarian distribution of space, and even the communal mosque does not show any element of social differentiation. Therefore, the existence of a larger free-standing residential building built using more detailed construction methods is somewhat discordant. Furthermore, the study of the faunal record recovered from the building indicates that the occupants’ consumption habits were privileged in comparison to the rest of the occupants in the enclosure. If a plausible scenario were to be suggested for the situation, the material evidence seems to indicate the existence of a hierarchy in the relationship structure between the volunteers. This verifies the presence of an officer or person in charge of the garrison and, therefore, indicates that the community mimicked the behaviour of regular troops. Please see Figure 4 below.

In turn, written sources describe various types of fortification along the coast of ancient Tarraconensis that could, hypothetically, be linked to the two main groups described: complex fortresses and spacious enclosures with no preserved archaeological structures or materials. Ibn Ḥayyān partially compiled a letter that the governor of Tortosa, ʿUbayd Allāh b. Yaḥyà would have written around 850–851 CE to Emir ʿAbd al-Raḥmān b. al-Ḥakam, informing him that owing to the taxes collected in his demarcation (min jibāyat ʿamali-hi) by his tax agents (li-ʿummālati-hi), he could rebuild his fortresses (ḥuṣūn) and pay the ransom for captives. Furthermore, he informed him that once the campaigns over the border against the Carolingians had ended, he had dispensed with the cavalry (miqnab [...] fursān al-ḥurs) sent from Córdoba, which he would have dispatched to the coastal enclosures (marābiṭi-him bi-l-māʾ) to continue guarding the coast (Ibn Ḥayyān 1973, p. 6). The subtle differentiation between fortresses and enclosures, as well as the governor’s indication to the cavalry to go precisely to the marābiṭ, enable us to finally suggest a probable distinction between the two terms.

The ḥuṣūn, which are well documented via cases such as Mollet, Marinet and, in particular, Tossal de la Vila, fit with the new constructions built from the end of the 8th century on the highest peaks along the coast and main communication networks. They were equipped with powerful defensive mechanisms and living quarters, which, in some cases, were somewhat complex and rather unique. In several cases, evidence has also been found of settlements that were located outside the enclosed perimeter, whose function is still being analysed. The difficult and almost impracticable access to these enclosures may also have been a premeditated choice, thus favouring the asceticism and isolation typical of the early spiritual expressions of Islam in the making.

Taking the case of the Tossal de la Vila fortress as a typical example, but also bearing in mind some of the characteristics from the Mollet and Marinet sites, its monumental and highly visible defences, as well as the distribution of living quarters along the inner perimeter of the wall seem to coincide with the parameters generally established for enclosures linked to the stable practice of ribāṭ. The existence in at least one of the enclosures of one of the oldest rural mosques in the Iberian Peninsula only strengthens this hypothesis, especially considering its marginal and peripheral location. A place of worship which, moreover, is consistent with the profoundly Islamic identity of the community settled there, who would have adopted early orthopraxis eating habits in which the consumption of swine would have been totally forbidden (Pérez-Polo et al. 2023a, p. 253; 2023b, p. 21).

The marābiṭ, whose etymology brings us closer to the root of ribāṭ and which lexicons are unanimous in linking to the concept of stables or enclosures where animals could shelter overnight, multiplied in slightly more accessible locations. They could be identified due to their location, which was along the same coastal communication routes and on the peaks of small mounts that were surrounded by thick walls built in the same way as the fortresses of the same period. However, the main difference was the absence of an internal structure to house groups of people on a stable and permanent basis. This suggests that the enclosures might have been used to temporarily house small mobile garrisons assigned to coastal surveillance, probably equipped with everything they needed to move quickly along the coast.

At this point, it seems coherent to suggest that the practice and the practitioners of ribāṭ could be the most relevant elements that were capable of building such a large number of contemporaneous fortified spaces with such homogeneous characteristics. Even more so if we take into account their location in a territory in which no rebellions or uprisings against the central power were documented during the Emirate phase, with defence being one of the main reasons for building fortifications during the period (Negre and Suñé 2019, p. 731). Moreover, the construction programme would have had the support of the local authorities, as highlighted by the fact that the governor of Tortosa himself used regular troops of his own or under his command to reinforce the coastal positions that generally belonged to the infrastructure of this institution.

All of the above leads us to conclude that, although they may have fulfilled different roles and were occupied by volunteer as well as regular troops throughout their existence, the drystone fortresses and enclosures located along the northern coasts of Sharq al-Andalus played a central and fundamental role in the practice of ribāṭ in this territory. The former as semi-permanent accommodation for various military groups and the latter as temporary staging areas for militarised mobile patrols. Their generalised, almost complete abandonment at the beginning of the 10th century only reinforces this approach, given that the success of the Islamic state led to a much more intense intervention on its borders, displacing these voluntary practices and entrusting the surveillance of the coast to regular troops and contingents. Their remote and inaccessible locations, while, at the same time, being manifest and visible to anyone travelling along the main coastal routes, combined the virtues required for the practice of militant spirituality yet also imposed restrictions that did not fit in with the new border defence parameters implemented after the proclamation of the Caliphate.

2.3. From the Practice of Ribāṭ to the Founding of a Ribāṭ

Without being able to certify whether it was first the practice of ribāṭ or the construction of sites to consolidate it, there is no doubt that the Ribāṭ Kashkī seems to have been one of the earliest cases of this kind of complex in the Iberian Peninsula. This is evidenced by the fact that it is mentioned by faqīh ʿAbd al-Malik b. Ḥabīb (d. 852), which enables us to establish its terminus ante quem around the first half of the 9th century (Al-Zuhrī 1989, pp. 90, 103; Negre 2015, pp. 116–20). Unfortunately, no material evidence remains of the building, which has been tentatively located near the municipal market in La Ràpita (Tarragona, Spain) based on the planimetries of the site prior to the Bourbon Reforms in the mid-18th century (Carles 2004, p. 143).

It seems unlikely, especially with no documented Viking expeditions against the Emirate’s coasts until 859 CE (Ibn Ḥayyān 1973, pp. 307–13), that the Northern European threat influenced its founding, as is documented in other cases. Neither do we have any information on large communities in Tortosa that developed heterodox practices during early Islam, another factor traditionally identified as a reason for its advent (Cressier 2019, p. 120). Even so, written sources attest to an episode in which a teacher (muʿallim) claimed to be a prophet and tried to spread an unorthodox interpretation of the Qur’an throughout the domains of Tortosa. The governor of the territory, the aforementioned ʿUbayd Allāh b. Yaḥyà quickly dealt with the situation and, after consulting with the ulema, decided to crucify the heretic in 851–52 CE (Fierro 1987, pp. 70–74) without any evidence to suggest that his preaching had spread or taken root in any major way.

Therefore, it is suggested that it might have been the personal initiative of an individual or group of people with strong ties to the site that played a relevant role in its founding, mainly in response to its location on both a land and a maritime frontier, as well as its role as a gateway to the frontier capital. In this regard, the main community that had ample presence on the Tortosa coasts and a fluid relationship with the religious currents linked to the practice of ribāṭ on the Maghrebi coast in the early 9th century was that of the baḥriyyūn (Negre forthcoming). In line with the explanations above, the origin of these seafaring communities on Tortosa has been well established, as well as their key role in the naval actions deployed from al-Andalus shores. However, it is also believed that these groups may have played a greater social role than they are usually given credit for and may have collaborated in the founding of the complex which concerns us here, with objectives and scope complementary to their own.

Written sources describe the Ribāṭ Kashkī as a place that, according to the visit that Al-Zuhrī himself claims to have made, had a water well whose level never dropped, even though thousands of people drank from it (Al-Zuhrī 1989, p. 103). This is undoubtedly one of the most recurring topos in stories within the genre of ʿajāʾib or ‘marvels’, which gave this type of building credence and a spiritual character marked by the blessing power (baraka) of a profoundly religious nature. Al-Idrīsī also mentions the place under the later name of Rābiṭa Kashṭālī. He comments on the solidity of its walls, which formed a fortified (ḥaṣīna) and impregnable (munīʿa) enclosure near a flourishing settlement with many buildings and houses (Al-Idrīsī 1866, p. 232). The importance this complex had for most Arab authors cannot be highlighted enough. Its core military and religious functions were not exclusive of each other or to the possibility of incorporating other complementary objectives that might be promoted from the location.

Although the most detailed descriptions of the building that have survived are from a later period, they all highlight the evident evolution of its objectives from its founding until the 12th century. This is similar to the transformation that can be observed between the ribāṭ and the rābiṭa of Guardamar, which progressively adopted a role primarily focused on its religious remit, which still endured after the feudal conquest. This is highlighted in a document signed in 1304, which stated that ‘pilgrimage to the place to where Muslims were accustomed should stop immediately’, although in 1327, the order was repealed (Bramon 2004, pp. 122–23). As a poignant epilogue, between 1610 and 1611, the site was used as a base of operations to embark and expulse the Mudejar and Moorish population from the Kingdom of Aragón and the lands of the Ebro, which came to a total of no less than 40,000 deportees from the local port of Els Alfacs (Colás 2010, pp. 18–19).

However, there is much we can infer about the originally intended functions of the Ribāṭ Kashkī. For example, it is important to highlight its role as a vehicle for the sacralisation process of the coastal areas and the dissemination of Islam itself, objectives inherent to the complex and in line with its religious matrix. This purpose was also attested to in the construction of other nearby buildings for worship, such as those already described in coastal fortresses. A fundamental mechanism in the configuration of the maritime and land frontier of al-Andalus, the place also flourished as a destination for fervent believers, such as the faqīh ʿAbd al-Malik b. Ḥabīb; the armed and spiritual defence of the frontier is a central part of its raison d’être. In short, it was a reception centre for believers willing to sacrifice their lives and worldly good to defend Islam and an opportunity for people from different backgrounds to make money from military campaigns. However, it is also a vantage point for the dissemination of the Muslim faith and the Islamic model of society.

In the same way that this dual military and religious function was paramount, its role in the specific tasks of surveillance of the territory would have been no less important. With the construction of the Ribāṭ Kashkī at the mouth of the river, together with the large contingent of ships under the command of the baḥriyyūn, both access to the city by river and the river crossing were effectively secured. It is also more than probable that the active coordination of the squadrons or units of volunteers, who would set out for the ḥuṣūn and marābiṭ along the coast, took place here. This suggests that there would have been an early warning system along the Tortosa coast that must have been tremendously effective throughout the 9th century right up to the proclamation of the Caliphate.

In keeping with this scenario, and as was confirmed centuries later when it became one of the Crown of Aragon’s most important ports, the site also acted as a strategic harbour from where many of the naval expeditions led by the baḥriyyūn, and later by the Emirate navy, would set sail. This role, due mainly to its natural characteristics, is later described by al-Idrisi in different mentions in his two major geographical works, in which he clearly indicates the role of the Rābiṭa Kashṭālī as one of the principal moorings (marsā) on the Tortosa coast (Al-Idrīsī 1866, p. 232; 1989, p. 161). A role that was clearly linked to coastal defensive and offensive actions against the Christians but which also had an intense connection to the commercial activity of these coastal communities, which included trading booty and prisoners captured during raids.

This leads us to conclude that the different groups of sailors and traders documented on the Tortosa coast must have played an important role in the founding of the Ribāṭ Kashkī. Their relationship with the Maghreb coast, where similar developments have been verified (Hassen 2001, p. 154; Fenwick 2018, p. 217; Cressier 2019, p. 120; Bosanquet 2023, pp. 6–7, 10–11; Van Staëvel 2023, p. 3), must be considered as one of the main channels of communication for the new spiritual currents along the al-Andalus coast. Given the evidence for the continuity of commercial imports as late as the 8th century in ports such as Tarragona, it would seem amiss not to acknowledge the role played by these communities in the transmission of ideas, knowledge and expressions of spirituality such as ribāṭ in the nascent Islamic world.

Naturally, the dynamism shown by the complex from its earliest beginnings was also reflected by the projection of new forms of peasant settlements which were built under its protection. In this regard, a large number of unpublished archaeological sites have been identified along the maritime coastline near the Ribāṭ Kashkī, most of which opted to settle in the vicinity of the complex. Judging by the pottery recovered from the site, the start of the colonisation process dates to the 9th century, although it experienced extraordinary growth throughout the Caliphate and Taifa periods (Martí and Negre 2023, p. 9).

This can be seen in greater detail in a document from 1097, which is an inventory of goods and properties of the institution in the context of a failed attempt by feudal lords to conquer Tortosa (CEU n.d., doc. 2). Here, the domains attributed to the institution in the 11th century included 29 populated locations (villulae) lacking ancient precedents, two-thirds of which would have been given gentilic names, such as Benifallim and Ventalles (Benjalima, Benalcale). These patronymic names may have corresponded to extended families as well as to settlement owners in the Late al-Andalus context. Both were small settlements with only a few houses or domestic units, which were mainly abandoned after the Christian conquest in the 12th century, evidenced by their onomastic disappearance. However, after the conquest, such neighbourhoods did not fit in with feudal logic, which preferred and encouraged compact populations, which led to the general abandonment of the old enclaves and the development of new densely populated settlements.

The fact that the territory flourished is due mainly to the role of the Ribāṭ Kashkī as a pole of attraction and a driving force behind new forms of settlement in an area that was largely depopulated prior to the 9th century. The growth in a number of settlements in the vicinity of the fortified enclosure and their integration into the agricultural strategies of the territory are also key to understanding the success of the enterprise. The possibility of establishing permanent relationships between the peasant communities who had settled there and the murābiṭūn and mujāhidūn who travelled periodically to the ascetic institution seems to have ensured the continuity of the latter. Please see Figure 5 below.

A small settlement known in sources as Codair, Alcozer or Alcover (CEU n.d., docs. 2, 5; CSCV n.d., doc. 849) would have originated in a place parallel to the Ribāṭ Kashkī itself. The name is related to the Arabic term al-quṣayr, which is a diminutive of al-qaṣr, which includes ‘public stopping place’ among its many different meanings (Franco 2005, pp. 42, 47, 51–52). It was once thought that the site had been established under the jurisdiction and control of the complex, but, in fact, it might have originated even earlier due to its strategic location at the crossing of the river Ebro, the main coastal road. Another of the strategic sites under the control of the Ribāṭ Kashkī, according to written sources, was Marmortua. This late toponym is usually used to describe salt mines, which the site would have had exclusive rights over before the feudal conquest.

Fishing also seems to have played an important role in the economic development of the region under the control of the Ribāṭ Kashkī and the settlements linked to it, at least since the 10th century. This is consistent with the fact that the district is located on the coast and the mouth of the river, an area where the quality and biodiversity of its waters are multiplied. Arab authors had already highlighted the important ichthyological richness of the region and, in particular, the lower course of the river Ebro where high quality, calibre mullets and sturgeons bred (Lévi-Provençal 1953, p. 103; Troyes and Reims 1883, p. 366). In this regard, fishing rights, which at the time belonged to the Ribāṭ Kashkī, were transferred by Count Ramon Berenguer III to the Monastery of Sant Cugat in his failed donation of 1097, thus confirming the importance of fishing and fishing rights in the region at that particular moment (CEU n.d., doc. 2). This can also be seen in the various lawsuits that seem to have passed between the Monastery of Sant Cugat and the Knights Hospitaller of Amposta after the conquest over that same issue (Beguer 1948, pp. 35–39).

From the above examples, it is clear that further to its initial military, religious and commercial objectives, the Ribāṭ Kashkī progressively acquired an important role in the administration and fiscal organisation of the territory, which later became a district within its own jurisdiction. This can also be inferred using the analysis of its complex, interwoven relationship with the economic sphere, which was manifested via its control of certain resources and services that were usually under state control. In this regard, the important role it played from the outset in the colonisation process of marginal and mostly unpopulated spaces, which it transformed into an emerging scenario full of vitality, should also be highlighted.

During the second half of the 9th century, the advent and consolidation of the Emirate’s navy, which was commanded by the state, had a profound impact on the activities the baḥriyyūn had performed for decades from the Tortosa coast, from where they had to flee first to the Maghreb and later to the southeast of the Iberian Peninsula (Negre forthcoming). With the coast protected by the new Cordovan fleet and an increase in hostilities on the Barcelona frontier, the most plausible scenario for the centre’s continuity was now based on the rise of asceticism and collaboration in the many aceifas against the Christians. This scenario would change even more significantly from the Caliphate period onwards, essentially due to the state’s new conception regarding the frontier, which involved major intervention from Córdoba (Negre and Suñé 2019, pp. 720–29). The deployment of large armed contingents to the nearby medina of Tortosa, the construction of new city walls, the building of shipyards to reinforce the Caliphate navy and the construction of a fortress at the mouth of the river all suggest that Córdoba would eventually assimilate almost all the military function that was once controlled by the Ribāṭ Kashkī (Negre and Martí 2015, p. 199).

Gradually, the ribāṭ became a rābiṭa, and the once multifaceted complex returned to its spiritual objectives, which, conversely, enabled it to continue to develop a complex process of colonisation and domination over its adjacent territory. In order for it to continue, from the Caliphate onwards, the complex had territorial and jurisdictional domains that put it on an administrative and economic par with the formal role that other territorial centres, such as the ḥiṣn, would play. However, given our current level of knowledge, it is impossible to determine what kind of internal organisation or authority, if any, this new centre would have had. If anything was evidenced by later written sources, it was that its role as a regional hub and place of pilgrimage and deep spirituality never became obscured.

3. Guarding the Sea, Tormenting the Enemy, Sacralising the Coast

Throughout this text, we have addressed the general discussion on the practice of ribāṭ that has been generated over the last few years based on case studies linked, primarily, to the eastern Mediterranean and the Maghreb. We have done so based on a new area of study that has provided complementary readings on processes, as well as a comparative analysis between scenarios. Northern Sharq al-Andalus presents unique and certainly complex characteristics that we believe modulate the practice of different expressions of both defensive and offensive armed spirituality. In this regard, its location as a dual maritime and land frontier, in the furthermost corner of Dār al-Islām, played a very important role in the social, economic and religious configuration of early al-Andalus. This was a framework in which the adoption of certain precepts of belligerent asceticism gave rise to the development of a fortified landscape mainly occupied by volunteers who were willing to comply with said precepts, but also served the military needs of a border authority that enjoyed a high level of independence.

In this regard, we have stated that this was the fate of a large number of believers who were eager to participate in a new religious scenario that elevated sacrifice in defence of Islam to one of the purest forms of the profession of faith, but also to other groups that saw ribāṭ as an opportunity to new and enriching endeavours. Under the name of murābiṭūn, mujāhidūn, muṭṭawwiʿa, and baḥriyyūn were very diverse groups with varying degrees of involvement in the practices performed, but all were committed in one way or another to the defence of the frontier and the spread of Islam. They comprised completely anonymous individuals from a wide variety of backgrounds, as well as renowned people, important fuqahā, ulema and other Muslim scholars. In general, they were participants in a permanent struggle who travelled to and from the frontier as if it were a pilgrimage, with the ultimate aim of actively defending the ummah by assaulting enemy positions, defending their own lands and providing logistical support for actions performed in relation to their objectives.

In the same vein, if anything can be inferred from the archaeological and written sources reviewed in this paper, it is that the practices linked to ribāṭ and jihād were not performed in a chaotic and disorderly manner but, in the case of Tortosa were planned meticulously. The evidence clearly supports this theory in the form of a series of fortifications and defensive enclosures that were built along the coast in the late 8th and very early 9th centuries that were capable of housing volunteer troops and, as already mentioned, accommodating regular troops who would have reinforced these positions in times of need. The ḥuṣūn and marābiṭ facilitated the permanent surveillance of the territory’s main maritime and land communication networks and were the material expression of the practices that would lead to the effective consolidation of an Islamic land in the making. These structural elements would, in turn, reinforce the visibility of Muslim power in the periphery while at the same time consolidating the presence of the new state and facilitating the dissemination of a series of norms, customs, traditions and economic patterns that would lay the foundations for a profound transformation of the Iberian Peninsula society. Sacralisation of the landscape via presence and visualisation in it.

It is worth highlighting the close ties between the governors of Tortosa and the practitioners of ribāṭ and jihād and, by extension, the fortified structure that had developed around them. Although we have already mentioned the role of governors as commanders of the regular and irregular troops established along the frontier, it is also important to mention that, in the mid-9th century, one of the governors communicated to the emir that with the income he had collected from taxes, he would take care of the ransom for the captives, rebuild the frontier fortresses and pay their officials. This highlights how the provincial authorities enjoyed special autonomy in the management of local resources during the period and a very close relationship with volunteer groups stationed there, as well as having use of the fortifications built along the coast.

On the other hand, the contemporaneous founding of a complex so unequivocally linked to these same practices as was the Ribāṭ Kashkī is extremely illustrative and only reinforces the sacralised and institutionalised character that these practices adopted. The founding of this complex had to take place at that precise moment, probably in response to the spread and consolidation of the practice of ribāṭ in the territory. These dynamics were also supported by the earlier presence at the site of a large community of baḥriyyūn, who were linked to seafaring activities that included jihād actions against Christian lands and ships and trading booty obtained from raids as well as transporting goods. The Ribāṭ Kashkī was founded, therefore, with the aim of centralising military and commercial activities that were already in place and supporting other constructions mainly linked to the practice of ribāṭ that had been built previously along the coast.

These foundational objectives were soon supplemented with other functions that the complex acquired as it consolidated its presence in the territory. Of particular note is its evolution as a judicial centre with broad territorial jurisdiction. The dynamics highlighted in this paper have already been proposed and argued for other similar enclaves throughout the Mediterranean; therefore, the case of Tortosa only seems to corroborate the strong ties between this type of fortified complex and their immediate environment as part of well-defined survival strategies.

Once the Caliphate was proclaimed, a major statisation of the frontier took place, which is reflected both in the appointment of its governors and in the troops mobilised for its defence. Tortosa began to be controlled by quwwād (sing. qāʾid) with an evident military character sent directly from Córdoba, whose powers, in many cases, went beyond the simple management of the territory. A paradigmatic example of this situation is the appointment of Mundir b. Saʿīd al-Ballūṭī, in 942 CE, as supreme authority of the frontier in Tortosa. His function was to supervise the judges and governors and inspect all those arriving from Christian territories (Ibn Ḥayyān 1979, p. 306). Along the same lines, governor ʿAbd al-Raḥmān b. Muḥammad b. al-Naẓẓām was appointed by order of the caliph and commissioned the construction of the city’s arsenal and shipyard, as well as the new city walls (Al-Ḥimyarī 1984, pp. 122–23).

This all points to the centralisation of frontier military action in the medina of Tortosa, where most of the regular troops, which now made up the bulk of the army, were garrisoned (Negre and Suñé 2019, pp. 728–29). Volunteers practically disappeared from the list of participants in campaigns and aceifas against Christian territory in this part of al-Andalus, with the regular army (jund), as well as the palatine guard (ḥasham) and other recruits (maḥshūdūn) taking centre stage. The study of the Islamic troops involved in the defence of Tortosa definitely suggests that, from around 920 CE onwards, Emir al-Nāṣir’s imposition over local authorities and the consequent increase in their tax revenues also facilitated the considerable increase in the activity of state troops in the territory.

Undoubtedly, this new situation, as documented, led to the definitive abandonment of the old ribāṭ places that were located on impregnable peaks along the coast and their progressive replacement by a dense network of state fortresses aimed at improving the organisation of the territories and the collection of taxes from positions closer to agrarian spaces (Negre and Suñé 2019, pp. 724–27). After more than fifteen years of pacifying every last corner of the Iberian Peninsula, the new caliph did not want impregnable barracks in remote, inaccessible places. This process must have been relatively rapid and culminated in eliminating both the ascetic spiritual-military practices and the buildings linked to them, at least in most cases.

However, one of the buildings survived. The former Ribāṭ Kashkī, now Rābiṭa, found its place in this new society of the victorious state by becoming an administrative and jurisdictional centre built around a religious institution that remained in force throughout the centuries to come. Associated exclusively with its spiritual functions and the support of pilgrims willing to undertake a spiritual retreat, from the 10th century onwards, its fate was tied to a completely Islamised frontier that was to see very different fortunes.