Introduction

In today’s world, where societies tend to be understood and assessed from a western standpoint, it is instructive to take a moment to reflect upon how dominant world cultures viewed the societies around them in earlier epochs. One such world culture was that of classical or medieval Islam which stretched from the Iberian peninsula to the borders of the Sinic and Indic worlds for a period which began in the late seventh century CE and continued well into the early modern era. During these centuries, a complex, sophisticated and universalizing civilization developed and flourished alongside the micro-cultures of this vast zone, which was known by Muslims as the dār al-islām, the house or domain of Islam. At the outset the dār al-islām was an Arab-dominated empire and Arab ethnocentrism led to the conceptualization of non-Arabs (^cajam), as inferior, both within and outside the empire’s boundaries. In subsequent centuries the dār al-islām was rarely a political unity in practice, despite the existence of the universal institution of the caliphate until 1258, and steady non-Arab conversion to Islam created a more varied Muslim community. Nonetheless, the dār al-islām came to possess normative linguistic, religious and cultural aspects which justified its inhabitants, Muslim and non-Muslim, in perceiving themselves as part of one world, and those outside as belonging to another, the dār al-harb, domain of war, which included both eminent civilizations such as China (Sīn), India (Hind) and Byzantium (Rūm), and less-developed ‘peoples’ (jins, ajnās).

The precise ways in which the inhabitants of the dār al-islām viewed the inhabitants of the dār al-harb were conditioned by a number of factors, which I shall mention briefly here. First, educated Muslims of the eighth century CE onwards, viewed their civilization as an heir to the great Near Eastern and Mediterranean civilizations of the past, including Greece, Rome, Levantine Byzantium and Persia. The great rupture which much Orientalist scholarship has perceived as occurring with the rise of Islam was not perceived as such by Muslims themselves, who understood their faith as a legitimate Abrahamic monotheism. They saw their culture as enriched by borrowing from their predecessors, despite the internal tensions between Arabs and non-Arabs. That being said, there was an alternative Muslim vision according to which the rise of Islam was indeed a decisive break with the preceding age of ignorance, the jāhiliyya, but this referred specifically to pagan Arabia, rather than to the Near East in its entirety. Second, Muslim appropriation of Greek learning, including Ptolemaic geography, provided a ‘scientific’ base for a worldview onto which religious considerations could be mapped. Third, Muslims in this period enjoyed the same sense of cultural superiority which Europeans and Westerners in general possess today, and this was reflected in their perceptions of the world beyond their frontiers, which brought together realistic eyewitness accounts and more impressionistic, ideological and, in some cases, fictive materials, portraying non-Muslims as an archetypal ‘other’.

The range of options Muslims possessed in viewing the ‘other’ can be seen quite clearly in the case of al-Andalus, the Iberian peninsula, significant areas of which were part of the dār al-islām from 711 until 1492. This region is particularly interesting due to its intimate and multi-layered relationship with the non-Muslim world, namely Latin Christendom and the pagan fringe of Europe. As a frontier territory, al-Andalus was one of the areas of the Muslim world which enjoyed fairly regular commercial and diplomatic contact with non-Muslim lands, and absorbed many individuals from outside the Islamic world as a result of war or purchase. It was also subjected to attacks from further afield, in the form of Viking incursions in the tenth century. Moreover, from the eleventh to fifteen centuries, Muslim territory was gradually lost and incorporated into the emerging Christian kingdoms of Portugal, Castile and Aragon, and ultimately modern Europe. This gives a particular piquancy to the task of presenting the region as part of a Mediterranean, but not necessarily European, world.

Various scholars have investigated Andalusi Muslim contacts with the north, and the images of Christians in Andalusi Muslim literature.Footnote ¹ Three main arguments emerge: first, the Muslims’ relative lack of interest in recording the political, cultural or social realities of life in neighbouring Christian lands, with which many were undoubtedly familiar; second, the tendency to use formulaic language rooted in religious difference, which positioned Christians as the enemy regardless of the actual situation; and third, a dependence on stylized tropes derived from eastern Islamic models, which give a Mashriqi colour to the literature. Whilst not the focus of this article, it is worth noting that Muslim writers showed equally little interest in the Mozarabs, the Christians living within Muslim al- Andalus.Footnote ² The main outlines of these arguments are incontrovertible, but it is possible to read often brief and stereotypical references to northerners ‘against the grain’ to elicit some sense of underlying realities. The incorporation of what might be termed ‘social comment’ and material evidence, alongside purely literary descriptions, has a similar effect.

The Andalusi experience illustrates the ideological rather than real nature of a tangible frontier between Christian Europe and neighbouring Islamic lands. Although Spain and Portugal have come to be viewed as part of Christian Europe, this geo-political and cultural ‘fact’ was by no means evident in earlier eras. The Iberian peninsula was part of the Roman empire, and became the site of a Christian Barbarian kingdom, that of the Visigoths, but its subsequent history belonged not only to Europe, but also to Islam. It was part of a Muslim trajectory, viewed as an integral part of the dār al-islām, and linked by religion, language and culture to North Africa and the Levant. One can posit ways in which the new Arab and Berber elite within al-Andalus saw itself as a legitimate heir to the preceding Visigothic aristocracy, and by extension to the Roman imperium. Such a view contradicts the reconquista narrative, which implicitly denied the existence of a substantial indigenous Arabic-speaking Muslim community in al-Andalus, and implied that the Arabs were a small conquering elite which was ultimately expelled, their temporary presence being no more than an Oriental aberration in Spanish Christian history.

In order to elucidate the ways in which Andalusi society positioned itself in relation to its Christian neighbours, I shall first look at broader Muslim views of the non-Muslim world, which provided the conceptual and literary framework within which Andalusis worked. I shall then address the situation in al-Andalus from three different standpoints: first, the competing ways in which the northern Christians and the new Arab Muslim regime viewed the Visigothic past; second, the ideological stance of the early Muslim governors, and then the Umayyads, towards Christian states and their diplomatic relations with them; and third, the role played by northerners of non-Muslim origin within Andalusi society itself. The main chronological focus of the article will be the period from the Muslim conquest in 711 until the fall of the Umayyad caliphate in the early eleventh century. This is because the subsequent conquest of the peninsula by dynasties of Maghribi Berber origin engendered a new set of circumstances, which necessarily impacted on concepts of political and cultural identity within al-Andalus.

Muslim taxonomies of space

In the realm of scholarship on the Islamic world, and in popular perceptions of Islam among both Muslims and non-Muslims, the concept of the division between the dār al-islām and the dār al-harb has taken on an exaggerated significance, which presents the issue in black and white terms that rarely correlated with actual practice. It is important to recognize that the terms ‘Islam’, meaning submission, and ‘war’ actually indicated the vertical relationship between man and God, not horizontal relationships between different human societies: Muslims had submitted to God, non-Muslims existed in a state of conflict with Him. Of course, warfare did occur across the divide, but practical exigencies frequently nuanced its bipolarity. Moreover, some jurists from al-Shāfi^cī onwards acknowledged an intermediate zone, the dār al- ul or dār al-ahd, the domain of truce, via which Muslims could maintain peaceful commercial and diplomatic relations with non-Muslims.

While Muslim scholars talked in such oppositional terms in theoretical legal and religious writings, the development of geography (and actual historical processes) in the early centuries of Islam indicated a variety of different taxonomies which were not entirely dependent on the religious division between the dār al-islām and dār al-harb, whether mediated by a ‘domain of truce’ or not. The contribution of Hellenistic thought, for instance, was enormous. Many Arab geographers mention Ptolemy among their sources, and the usual arrangement of materials followed a Greek-derived rather than Islamic schema. This divided the world into seven climes inhabited by various peoples, or human groups, whose characteristics reflected the environment in which they lived, in accordance with humoral theory.Footnote ³

The educated Muslim view of human geography in the classical Islamic era is perhaps best explicated by the polymath, al-Mas^cūdī, a rather independent intellectual, who wrote on a wide range of subjects, of which geography was just one. Abū’l-asan ^cAlī b. al-usayn al-Mas^cūdī was a native of Baghdad whose nisba, al-Mas^cūdī, was said to indicate his lineage’s descent from Mas^cūd, one of the companions of the Prophet.Footnote ⁴ His exact birth and death dates are unknown, but it is generally thought that he was born between 893 and 898, and that he died in 956 or 957.Footnote ⁵ He travelled extensively, and wrote historical- geographical works from personal knowledge as well as received wisdom. The most famous are Murūj al-dhahab wa-ma ^cādin al–jawhar (The fields of gold and mines of gems) and Kitāb al-Tanbīh wa’l-ishrāf (The book of warning and oversight).

A detailed analysis of these texts is beyond the scope of this work, but, as Ahmad Shboul has pointed out, one of the characteristics of al-Mas^cūdī’s oeuvre is his incorporation of information about the world beyond the dār al-islām, in both a chronological and spatial sense. Shboul discusses al-Mas^cūdī’s treatment of pre-Islamic and extra-Islamic history, notably Byzantium and the empires of China and India, in addition to less well defined groups such as sub-Saharan Africans (Sūdān), the Turkic peoples (ajnās al-Turk) among whom he includes the Khazars, the Rūs, the Slavs (Saqāliba), the Bulgars (al-Bulghar), and the Franks (al-Ifranja), as well as the Galicians (al-Jalāliqa) and the Basques (al- Washkansh).Footnote ⁶ What is distinctive about al-Mas^cūdī’s information is its level of detail, especially with regard to the Franks and Slavs, both groups of significance to the Muslims of al-Andalus. Al-Mas^cūdī’s access to and interest in unique materials, such as his list of Frankish kings, may have been what led some of his successors to assume that he was of western Islamic origin.Footnote ⁷ This was not the case, but the document itself was of Andalusi provenance, and it reflected tenth-century Andalusi interest in lands to the north of the Iberian peninsula.

In this context it is worth noting that, despite my decision to regard al-Andalus as external to Europe in the later sense of the term, some Muslim geographers did identify a geographical entity, Arūfā, which contained al-Andalus (Iberia, the Vandals) al-Saqāliba (Slavs), al-Rūm (Byzantium, the Greeks and possibly Latins) and Firanja (Francia, the Franks).Footnote ⁸ However, since these terms denote both peoples and their territories, it is not clear whether the Muslim inhabitants of al-Andalus were assimilated in this model, or perceived to be outside it. Other sources described the landmass abutting al-Andalus as ‘the great land’ (al-ar al-kabīra), a term sometimes taken to mean Europe but, in fact, extremely vague in terms of its boundaries, and devoid of religious or cultural content.Footnote ⁹

Who were the heirs of the Visigoths?

My starting point in this attempt to elucidate Andalusi Muslim attitudes towards the north is to consider how Muslims during the Umayyad era viewed the land and people they had conquered, both of which were appropriated by northern Christians, the Asturians in particular, as parts of Latin Christendom temporarily lost. The myth of the reconquista, and the consequent depiction of Muslims as foreign interlopers in Christian Spain, began to emerge as early as the ninth century at the Asturian court of Alfonso III (866–911). The anonymous author of the Crónica de Albelda presented a historical trajectory, which moved from Rome to Byzantium, to the Germanic-Gothic kings whose kingdom was occupied by the Saracens, and then created an entirely fictive link between the Visigoths and the Asturian kings, whom he depicted as engaged in a battle to regain ‘Spania’ for Christianity.Footnote ¹⁰ His work was based in part on the Crónica Profética written by a Mozarabic Christian, who had migrated north from Cordoba in the wake of the mid-ninth-century Martyr Movement, during which forty-eight Christians were put to death by the reluctant Umayyad authorities for blasphemous statements and apostasy. The Crónica de Alfonso III expressed a similar viewpoint. In these texts, Muslims were described using a vocabulary which Manzano Moreno describes as ‘classical and biblical’Footnote ¹¹ but which was also orientalizing and, as Barkai points out, inherently more hostile than earlier extant Christian narratives.Footnote ¹² In this manner, the Christian kings of Asturias and its successor, the kingdom of León, presented themselves as the heirs of the Visigoths and the Muslims as the ‘enemy’.

Ironically, the Goths of Septimania, a region which formed part of the Upper March between Muslim and Frankish territory either did not claim this inheritance, or expressed it from within the Muslim community. For instance, Mūsā b. Mūsā of the Banū Qasī lineage, which claimed descent from the Visigoth Casius and held the city of Tudela, is said by the Crónica de Alfonso III to have described himself as the third king of Spain, alongside the rulers of Oviedo and Cordoba.Footnote ¹³

Alongside the vocal and familiar claims to the Roman and Visigothic heritage made by the Asturians, one must place the weight of Muslim Umayyad historiography, in which religion was not a barrier to the Muslim inheritance of non-Muslim monarchical and imperial traditions. The purported title of Mūsā b. Qasī emphasizes this point, but it was also latent in the geo-political definition of Muslim al-Andalus, the relations between the Umayyads and the Visigoths, Umayyad literary narratives, and even urban planning. Although Muslim references to the Roman or Visigothic past were often implicit rather than overt, it is clear that the rulers of al-Andalus in general, and the Umayyads in particular, envisaged themselves as its heirs. This was in keeping with the attitude adopted by the Umayyads in Syria towards the Byzantine imperial heritage, and the ^cAbbasids’ extensive interest in the intellectual production of the ancient world, manifested in the translation of numerous works from Greek to Arabic in Baghdad between the eighth and ninth centuries.

In the geo-political sphere, the extent of Muslim-held territory coincided with the area controlled by the Visigoths and the Romans before them, both of whom had experienced significant difficulties in controlling the tribal peoples of the mountainous northwest of the Iberian peninsula, the Vasconi, the Asturi and the Cantabri, whose conversion to Christianity only began in the sixth century. The Arab commander, Mūsā b. Nuayr, who initiated the Muslim conquest of Galicia in 714, satisfied himself with occupying the fortified sites associated with the Romano-Visigothic limes and subsequent Muslim conflicts with the northern mountaineers continued a pattern in place for centuries.Footnote ¹⁴ Whilst the inhospitable mountainous terrain of northwestern Iberia is frequently cited as the cause for the Muslim failure to ‘complete’ the conquest, their relative willingness to enter the Pyrenees suggests that it may not have been the terrain alone which deterred them, but also a recognition that they had reached a natural political limit.

The little that we know about Muslim interactions with the Visigoths from historical and literary sources suggests a similar continuity. According to the Chronicle of 754, āriq’s rapid victory against the last Visigothic king, Roderic, was facilitated by the machinations of Oppa, son of Egica, a previous Visigothic ruler, who assisted the Muslims. Within a few years of the conquest, there is evidence of agreements and alliances being struck between Muslim commanders and Visigothic notables. The best-known example of such an alliance is the Treaty of Theodemir or Tudmīr, a document negotiated by ^cAbd al-^cAzīz, son of Mūsā b. Nuayr, which allowed the Visigothic dux of Murcia to retain control of his lands and his people to practise Christianity, in return for a stipulated tribute in kind.Footnote ¹⁵A similar deal was made around 714 with Casius, the eponymous ancestor of the Banū Qasī lineage mentioned above, which controlled the upper Ebro valley from Tudela until the first quarter of the tenth century. Such alliances were frequently followed by the conversion of such notables, or their descendants, to Islam, thereby creating a Muslim Visigothic strand within the new ruling elite.

This mapping of one elite onto another, or their merging, was further fostered by inter-marriage. ^cAbd al-^cAzīz, son of Mūsā b. Nuayr, who became governor of al-Andalus on his father’s departure for Syria in 714, married Ayla, the widow (or daughter) of the last Visigothic king, Roderic, whose desire to see her new husband as a monarch in the Visigothic mode rather than as an Arab governor led to his assassination.Footnote ¹⁶ More significant perhaps is the story of Sara, granddaughter of King Witiza and ancestress of the famous Umayyad historian Ibn al-Qūiyya, whose name (son of the Gothic woman) probably commemorated his royal Visigothic lineage. According to the historical narrative presented by Ibn al-Qūiyya, Sara’s place in the new Muslim order was granted to her by the Umayyads of Damascus, upon whose mercy she had cast herself. Out of respect for her royal lineage, they treated her as a member of the family and found her a husband.Footnote ¹⁷ When she was later widowed in al-Andalus, her second husband was selected by ^cAbd al-Ramān, the first Andalusi Umayyad ruler, further emphasizing the link between the two royal lines.Footnote ¹⁸ Although one may argue that the Arab emphasis on patrilineality made this transfer impossible, female lines could and did figure in the transmission of symbolic capital, the importance attached to descent from the Prophet’s daughter, Fāima, being a case in point.

This link was naturally enhanced by the collapse of the Umayyad caliphate of Damascus in 750, ^cAbd al-Ramān the Umayyad’s escape to Iberia, and his creation of an independent Iberian-based polity coterminous with the previous Visigothic kingdom. Whilst Umayyad legitimacy was overtly linked to their descent from the third Rightly Guided Caliph, ^cUthmān, and their status as ‘sons of caliphs’ in keeping with Muslim paradigms of power, the Romano-Visigothic past also figured in Umayyad ideology. In his study of Umayyad historical narratives, Gabriel Martinez-Gros shows how tenth-century historical-literary texts, such as the anonymous Akhbār Majmū ^ca and the Chronicle of Ibn al-Qūiyya, present narratives which depict Iberia as a promised land. This land was opened up and to an extent offered to the Muslims by its rightful heirs, the family of Witiza, whose royal position had been usurped by Roderic.Footnote ¹⁹ Moreover, the emphasis placed on Mūsā’s conquest of Seville, ‘la capitale de l’Espagne romaine’, and Mérida, ‘dont les prestiges antiques sont longement soulignés’, highlighted the fact that the Umayyads did not only inherit the legacy of Witiza but also that of the Romans.Footnote ²⁰ A similar narrative underlies Ibn ayyān’s treatment of ^cAbd al-Ramān III’s subjugation of the city of Toledo, which he frames within a discussion of Toledo’s earlier glories as a Roman and Visigothic centre, and its propensity for rebellion, both of which are drawn together in ^cAbd al-Ramān’s triumphant reintegration of the ‘mother of the cities of al-Andalus’ into the Umayyad body politic.Footnote ²¹

Andalusi Umayyad architecture and urban planning are another domain where subtle accommodations with the Roman and Visigothic past may be detected. It has frequently been noted that the Umayyads showed a certain nostalgia for their Syrian homeland, and that their architecture made multiple references back to the monuments of the first Umayyad caliphate in Syria. However, ^cAbd al-Ramān I also adapted local Visigothic custom, a politically wise move, given that the majority of the population was indigenous and Christian until well into the ninth century, and therefore not susceptible to Islamic or Syrian spatial symbolism. The symbiosis achieved between Visigothic and Umayyad religio-political planning is most evident in the placing of the great mosque and palace in Cordoba, both of which were (re-) constructed in the 780s. Umayyad urbanism in the Middle East was frequently distinguished by the placement of the ruler or governor’s residence, the dār al-imāra, adjacent to the great mosque. In the mature Umayyad model, the governor’s residence was actually attached to the southern qibla wall of the mosque.Footnote ²² In Cordoba, however, ^cAbd al-Ramān I adapted this model to reflect the former Visigothic disposition of space: he constructed his residence on the site of the Visigothic palace, already used as a Muslim governorial residence, and the great mosque across a street to the east on the site of the church of St. Vincent.Footnote ²³ He thus created a new urban core, which could be read by the Arab Muslim elite as evoking Umayyad precedent, and by the Christian subject population as a continuation of the Visigothic tradition.

The Umayyads’ appreciation and appropriation of the imperial Roman heritage was manifested in their careful restoration of the Roman bridge across the Guadalquivir river in Cordoba, which numerous Umayyad and post-Umayyad chroniclers identified as one of the wonders of Islamic Cordoba.Footnote ²⁴ The architecture of the great mosque of Cordoba also made reference to Roman constructions in al-Andalus, whilst simultaneously evoking Syrian Umayyad precedents. Although the use of double-arcades had already occurred in the great mosque of Damascus and the al-Aqā mosque in Jerusalem, the hallmark red and cream voissoirs of these arcades in the great mosque of Cordoba echoed the design of the Roman aquaduct at Mérida, a recognized site of antique splendour, as mentioned above. Finally, ^cAbd al-Ramān I’s construction of a suburban estate called Ruāfa, after the Umayyad country estate on which he had grown up in Syria, may be seen as ‘nostalgia’, but it was equally mapped onto the Visigothic and earlier Roman practice of maintaining villas in the hinterland of Cordoba.Footnote ²⁵

It therefore seems legitimate to consider the Arab conquerors of al-Andalus and the Umayyads as the heirs of the Visigoths in both real and imagined ways. The Iberian peninsula entered the Muslim consciousness as part of a Mediterranean imperial world, which had been inherited first by the Visigoths and then by the Muslims. The religious difference between conquerors and conquered did not create an insurmountable barrier to the integration of elements of the Visigothic elite, including the monarchy, into the new Arab Muslim elite which had assumed power. In fact, the two groups shared a community of interest. The sustained opposition which the conquerors faced in the northwest was not evidence of Visigothic or Christian resistance to an oriental Muslim invasion, but the continuation of older patterns of resistance to the authority of the south. This gained a new focus by the Asturian appropriation of the Visigothic past in a strident, but not historically veracious, way.

The Umayyad view of the north

The Andalusi Umayyad worldview thus posited the existence of a Mediterranean world, which correlated with the climes most conducive to civilization, as identified by Muslim geographers. Their attitude to the lands and peoples to the north was partially predicated, therefore, on their location in less favourable climes. Their religious affiliation, Christian or pagan, added another layer to theoretical discourses of difference based on the binary distinction between dār al-islām and dār al-harb. However, the very real connections across ethnic, linguistic and faith boundaries which existed in the northern marches (thughūr) created a reality which often contravened textual bi-polarity. It is perhaps helpful to consider the ‘north’ in terms of three distinct but overlapping zones: first, the duchies, principalities and kingdoms of northern Iberian and the Pyrenees, with which Muslims, particularly in the marches, had intimate and regular contact; second the Frankish and Germanic empires, with which there were diplomatic and commercial connections of fluctuating significance; and third, the semi-pagan fringe of Europe inhabited by Vikings (Majūs) and Slavs (Saqāliba), about which almost nothing was known, and which thus enjoyed a somewhat fictive status in Andalusi consciousness.

Before surveying the sources, however, it is necessary to sound a cautionary note, for relations with non-Muslims in northern Iberia and beyond were not of primary significance to Muslims prior to the eleventh century. Al-Andalus was oriented towards the Maghrib and the Middle East, rather than northern Europe. Although the northern marches (thughūr) were the site of frequent, sometimes annual, military activity, many engagements considered major by the Latin sources, are barely mentioned, or described in a highly formulaic manner, on the Muslim side. Moreover Muslim chroniclers show relatively little interest in the diplomatic exchanges which did occur between Christian territories and the Umayyad state, which are recounted in much less detail than the comparable visits of Berber chiefs from North Africa. Although certain events, such as the defeat of ^cAbd al-Ramān III by Ramiro II of Léon at the Battle of Alhandega (al-Khandaq) in 939, receive extensive coverage, information on the north is primarily given in terms of its relevance to the politics of al-Andalus, rather than its intrinsic interest or significance.

In discussing relations between the Muslims of al-Andalus and their immediate northern neighbours in Iberia and southern France, the Latin and Arabic sources both draw on bi-polar meta-narratives. On the Christian side, Latin chronicles, such as the Histoire générale de Languedoc, quickly evoked a Christian versus Muslim logic. The Mozarabic chronicle of 754 talked of Europeans (Europenses) fighting the ‘Saracens’ or ‘Ismaelites’ in the context of the Battle of Poitiers (balā al-shuhadā’).Footnote ²⁶ Muslim sources used a similar generic vocabulary, which included terms such as enemy (^cadū), or lands of the enemy (bilād al- ^cadū), polytheists (mushrikūn), and infidels (kuffār), along with the purely descriptive terms for Christians (rūm, na ārā). A pejorative term quite frequently used was ‘renegade’ (^cilj) implying not only religious wavering, but also potentially servile status. For instance, Ibn ^cIdhārī repeatedly describes the population as ‘renegades’ during his account of the Muslim conquest of Iberia, and then later uses it for northerners such as Sancho I of Navarre and his subjects.Footnote ²⁷ Another word used for northern rulers was ‘tyrant’ (āghiya) which meant an illegitimate rather than bad ruler in Muslim religio- political discourse.

The perspective or narrative thread produced by the repeated use of this terminology was that the natural state between Muslims and Christians was one of war, conceptualized as a holy war (jihād) on the Muslim side, and as the reconquista on the Christian side. The most obvious employment of this semantic complex occurred in the Muslim sources in descriptions of the annual, usually summer, military campaigns known as awā’if (sing. ā’ifa). These campaigns were not always directed against non-Muslim territory, and could equally serve to quell provincial rebellion or unrest. However, when they were directed against the Christian north, the language of jihād was a sine qua non. To take a random example, in Ibn ayyān’s report of the ā’ifa of 924 directed against the Basque country, he inserts the phrase ‘may God destroy it’ (dammarahā Allah) after mentioning the destination, and then states that the army travelled through the borderlands of the ‘polytheists’ (mushrikīn) and raided and burnt the ‘land of darkness’ (ar al-kafra).Footnote ²⁸ Such ideological-religious posturing was accompanied by more practical reportage. The commander of the ā’ifa wrote to ^cAbd al-Ramān III to inform him of the opposition that Ramiro II of León (Rudhmīr) faced from two lineages, described as the Banū Gomez and Banū Ansurez. In league with his brother, Alfonso, they had plunged the country into near civil war, to the advantage of the Muslims.Footnote ²⁹ Although many ā’ifa entries are brief and formulaic, they form the sole entry for some years, and thus appear as part of the framework of the narrative itself, highlighting the conceptual importance of jihād to Umayyad political self-identity by the tenth century, if not before.Footnote ³⁰

However, the majority of these sources also hint at less conflictual attitudes, which contradict their respective meta-narratives predicated on the concept of the existence of a strict frontier between the dār al-islām and Christendom, and show the complexity of the relationships which developed in the marches. Despite the tendency of Muslim chroniclers to depict the marches as the site of an ongoing jihād, alliances frequently cut across confessional lines, indicating an environment where local lords shifted their allegiance between Carolingian Aix-la-Chapelle and Umayyad Cordoba on purely political grounds. The Muslim advance through Septimania towards Francia, in the first half of the eighth century, appears to have been facilitated by Gothic and Aquitainian support for the Muslims against the Franks, and it is possible that at least one Gothic rebel against the Carolingians offered his allegiance to the Umayyads later on.Footnote ³¹ Conversely, the Carolingian acquisition of Gerona (785) and Barcelona (801) owed much to the support, albeit rather wavering, of various Muslim opponents of the authority of the Umayyads in Cordoba.

The Iberian Muslim governors and notables of Mérida and Toledo displayed similar tendencies. In 854–55, the Toledans sent a message to Ordoño I requesting Asturian assistance against the Umayyad ruler, Muammad I, who was approaching the city with an army.Footnote ³² Conversely, in times of internal strife in the Christian kingdoms, either side could seek the support of Muslim Cordoba. One interesting exchange of this kind began in 956, when ^cAbd al-Ramān III responded positively to a request from Ordoño III of León for a truce (ul ).Footnote ³³ In 962, however, Sancho I, Ordoño’s brother who had succeeded him, sent messengers with the qā ī (judge) of Valencia and other Muslim notables to offer his oath of allegiance (bay ^ca) to the new Umayyad caliph, al-akam II, in return for support against his cousin, another Ordoño.Footnote ³⁴ Inter-marriage further maintained relationships across religious lines. The Banū Qasī lineage, whom I have mentioned several times, enhanced their power and prestige in the Upper Ebro valley through inter-marriage with the Christian lords of neighbouring Navarre. The Umayyad amirs themselves showed a preference for Basque and Navarrese wives and concubines, and at a later date Alfonso VI of Castile had a relationship or marriage with Zayda, a woman described as the daughter-in-law of Mu^ctamid of Seville, and possibly the daughter of the ā’ifa king of Denia.

It was primarily the unstable political situation in the marches, and the absence of any clear border, which brought the Muslims of al-Andalus into diplomatic contact with Frankish rulers. Other contacts, of a commercial nature, were fostered by merchants, who usually operated independently. Although Franks (afranj) was sometimes used in a broad way by Muslim authors, Andalusis tended to identify them more specifically as those peoples subject to the Merovingian and then Carolingian kings. The geo-political proximity of al-Andalus to the Frankish realms made this a more developed relationship than the later contacts made with the Ottonian realm. As the previous paragraphs on relations with northern Iberia show, the role of dissidents, Christian and Muslim, was crucial to the formation of diplomatic relations, in which religion played little overt part. Possibly from 788, and certainly by 810, the Umayyads and Carolingians had begun to exchange embassies to negotiate truces. The immediate cause was the sustained opposition of the Umayyad brothers, Sulaymān and ^cAbd Allah to the accession of their brother al-Hishām, followed by his son, al-akam, and their attempt to elicit Carolingian support. In the multiple reports of ^cAbd Allah’s visit to Charlemagne, there is no trace of the generic hostility expressed in other contexts, for the language is neutral and descriptive. Francia is simply called ifranja or bilād al-faranja (land of the Franks) and Charlemagne is described as malik al-faranja (king of the Franks or Francia).Footnote ³⁵

These early mentions of the Franks are expedient, and do not demonstrate appreciable interest in northern territories based on intellectual curiosity. This attitude appears to have shifted during the tenth century, after the Umayyads assumed the caliphal title of Commander of the Faithful (amīr al-mu’minīn), and constructed the royal city of Madīnat al-Zahrā’ outside Cordoba, as a stage for ceremonial intended for not simply a domestic but also an international audience. A small indication of Umayyad interest in the politics of the north is al-Mas^cūdī’s reproduction of a summary history of the Frankish kings, said to have been presented by Bishop Godmar of Gerona to the Umayyad caliph al-akam II (961–76) some years before his accession. Al-Mas^cūdī discovered the volume in al-Fusāt in Egypt, and included it in his geographical-historical work, the Murūj al-Dhahab. There is other evidence of a broader Umayyad interest in the outside world in this era. Diplomatic exchanges with Byzantium led to the translation into Arabic of copies of Dioscorides’ De materia medica and Orosius’ Historiae adversum paganos in Cordoba. A sense of enquiry encouraged individuals, such as the eminent Jewish courtier and physician asdāy b. Shaprū, to make contact with the Jewish ruler of Khazaria on the Volga.Footnote ³⁶

One may see the contacts which developed with the Germans during the second half of the tenth century as part of this new interest and awareness. Contact began with a series of embassies between the courts of Otto I and ^cAbd al-Ramān III in the 950s, followed by a later mission to the court of al-akam II. Our fullest account of this series of exchanges is located in the biography of the head of the second Ottonian mission, Jean de Gorze.Footnote ³⁷ The embassy which Recemundo, Bishop of Cordoba, led to the court of Otto on ^cAbd al-Ramān III’s behalf is unfortunately not recorded in the same detail. According to Jean de Gorze, one of Otto’s motivations in initiating the exchange was to persuade the Umayyads to force the Muslims of Fraxinetum, a small enclave north of Marseille, to desist from piracy and brigandage in his southern domains.Footnote ³⁸ Nonetheless, a certain amount of religious polemic seems to have been evident in the correspondence of both sides, and Jean de Gorze certainly viewed his mission as one which might require martyrdom.Footnote ³⁹ The tone of his biography wavers between the two registers identified above. On the one hand, al-Andalus is called ‘Spain’, the caliph is called ‘king of Spain’ and its Christian inhabitants at least are ‘Spanish’. Recemundo is called a Spaniard and described as ‘a good Catholic, perfectly educated in our [Latin] culture as well as that of the Arabic language’.Footnote ⁴⁰ On the other hand, the initial description of ^cAbd al-Ramān suggests both violence and disgust. The biography states, ‘that king, a sacrilegious and impious Saracen, and as such an absolute stranger to the true faith, at the same time as soliciting the friendship of a Christian prince, had vomited blasphemies against Christ in the letter which he had sent’.Footnote ⁴¹

Despite such mutual assertions of religious superiority, both sides were apparently willing to temper their language in recognition of the lack of real political competition between the two realms. In the description of Jean’s two meetings with ^cAbd al-Ramān III, it seems that the Umayyad caliph viewed the encounter not as an embassy or audience in the usual sense of the word, but as a more informal meeting between himself and a non-Muslim holy man, who offered wisdom and knowledge, as well as practical information about the Ottonian polity. The conversation itself, as reported in Jean’s biography, reveals its practical rather than polemical nature: ^cAbd al-Ramān III asked about Otto I’s power and wisdom, his military capacities, his wealth, his knowledge of the arts of war and his victories. When Jean asserted Otto’s superiority in all these respects, ^cAbd al-Ramān countered with an example of the former’s political imprudence, his failure to monopolize political power for himself. This comment suggests some foreknowledge of politics in Saxony, derived perhaps from Recemundo’s visit, which occurred whilst Jean was in Cordoba.Footnote ⁴² Daniel characterizes the diplomacy of the era as ‘complimentary’ and ‘exploratory’ in nature, a description which Barceló finds particularly apt for the exchanges between the Umayyads and Ottonians, which reflected their imperial aspirations rather than specific objectives.Footnote ⁴³

The missions which arrived during the reign of al-akam II, between 971 and 974, are the subject of conjecture as to the identity of the Frankish or German rulers who sent them, Hugh Capet or Otto II, and their date. Neither appears to have elicited substantial Umayyad interest, given the lack of clarity in the Arabic sources. The only extant indication we have of further ‘academic’ interest in the north is the fragmentary travel account of northern Europe, written by the Andalusi Jew, Ibrāhīm b. Ya^cqūb al-urūshī, in the 960s. Al-urtūshī was a private scholar and traveller, who met the Holy Roman Emperor, Otto, in an informal audience during his travels. The identity of al-urūshī, his itinerary, and the purpose of his journey are all debated, but his ri la describing Germanic and Frankish lands provided important material for other geographers, notably al-^cUdhrī and al-Bakrī.Footnote ⁴⁴

The world beyond Frankish and German territories was the most distant from Andalusi experience, and barely entered into their imagination before the Viking raids of the mid-ninth century. In 844 the governor of Lisbon drove away eighty Viking long boats, which then travelled up the Guadalquivir river to raid Seville.Footnote ⁴⁵ Further Viking raids occurred in 859 and 966, with additional sightings in 971.Footnote ⁴⁶ The standard term adopted for the Vikings in Andalusi chronicles is ‘Majūs’, originally used in the Middle East for Zoroastrians, but employed in North Africa for pagans in general.Footnote ⁴⁷ The association appears to have been due to the erroneous perception that both Zoroastrians and Vikings were fire-worshippers.

At the time that the events occurred, Muslims knew very little about the Vikings and other peoples of the far north, but the usual term which came to be employed in the Islamic east was Rūs, which overlapped with Saqāliba on occasion. Ibn Khurradadhbih, writing before 850, mentions the Rūs as traders in furs and swords, one of whose routes lay through Francia or al-Andalus and North Africa. In the work of al-Ya^cqūbī (891), the Rūs and Majūs are explicitly equated in his brief account of their attack on Seville in 844. More detailed notices appear in the form of the ri la (travel account) of Ibn Fadlān who accompanied an ^cAbbasid caliphal delegation to the King of the Volga Bulghars and Saqāliba in 912, and the tenth-century geographical works of al-Mas^cūdī (d. 956) and others.Footnote ⁴⁸ Andalusi writers of the tenth to eleventh centuries continued to describe the Vikings as Majūs.

The lack of clarity in the Arabic sources concerning the identities of Vikings and Slavs reflects the lack of real engagement with them, engendered by their distance from the dār al-islām and their fleeting contacts with it. This was particularly true for the inhabitants of al-Andalus. Although Scandinavian and Russian coin deposits contain numerous coins from the Muslim east, they contain very few Andalusi coins. According to Barceló, those that did reach Scandinavia appear to have come via the Middle East, the Caucasus and Russia, mixed in with Carolingian coins, and perhaps from Viking raids on western al-Andalus.Footnote ⁴⁹ He further contends that coin deposits are not a necessary indication of commercial activity. The Vikings therefore entered the Andalusi consciousness first and foremost as a sudden, mysterious, military threat. The sporadic and hostile nature of their incursions precluded the development of extensive contacts, and they came to be incorporated into the Andalusi worldview as a literary ‘other’, throwing into relief the normative characteristics not simply of Muslims but also of the inhabitants of civilized Mediterranean climes.

This is most apparent in the account of the mission of the poet-courtier, al-Ghazāl, to the Majūs, which was supposed to have taken place after the 844 raid. The choice of al-Ghazāl as ambassador was predicated on his appointment on a mission to the Byzantine emperor, Theophilus, in 840, a mission which is alluded to in the Muqtabis of Ibn ayyān.Footnote ⁵⁰ However, al-Ghazāl’s mission to the Vikings is only mentioned several hundred years later, by the Andalusi scholar Ibn Diya (1155–1235) in his compilation of western Islamic poetry. This has generated lively debate as to the authenticity of the account. Levi-Provençal rejected it, while al-Hajji made a determined attempt to argue for its core veracity despite its formalized literary form.Footnote ⁵¹ The most recent scholarship on the matter suggests that the account is indeed fictional.Footnote ⁵² However, this does not prevent it being based on a commonly held image of the Majūs, their culture and their lands in a similar manner to the satirical ri la of Abū Dulāf, composed at the Samanid court of Bukhara.

According to Ibn Diy?a’s account, which al-Hajji presents in Arabic and an English translation,Footnote ⁵³ the Majūs inhabited a series of islands and the mainland nearest to them. They were Christian, apart from a minority who remained fire-worshippers and were subjected to attacks and enslavement by their Christianized kin. Their women were highly independent and, indeed, it is the Majūs queen’s open flirtation with al-Ghazāl, and her assertion of her right to choose male partners, which provides the focus of the account and titillation for the reader. A parallel thread, however, is al-Ghazāl’s sophistication and superior intelligence in his encounters with the Majūs rulers. In a passage highly reminiscent of ^cAbd al-^cAzīz’s attempt to assume Visigothic monarchical pretensions by making his fellow Arabs bow before him, the king of the Majūs is described as making the entrance to his presence so low that al-Ghazāl must kneel to enter. He avoids this by shuffling through on his bottom with his feet in front of him, thus showing the king the soles of his sandals rather than genuflecting. This sense of superiority is also apparent in al-Ghazāl’s dealings with the queen whom he successfully flatters but does not succumb to.

‘Northerners’ in al-Andalus

The final aspect of the relationship between al-Andalus and the north under consideration here is the use of northern European slaves and captives within Andalusi society, as court eunuchs, soldiers, and, in the case of women, concubines and sometimes wives. Whilst the majority of our evidence comes from the Cordoba-based chroniclers of the Umayyad period they do also make reference to the broader social impact of these individuals on Andalusi society. Both ethnic and status or service oriented terms were employed to designate them. For males, the most common ethnic term was ‘Saqlabī’ (Slav), followed by ‘Frank’, while fatā (youth) was a common status designation which indicated their dependency on their master, and their lack of full social maturity. A frequent term for females was ‘Rūmiyya’, in the sense of ‘Christian’, and the majority entered the historical record as individuals when they gave birth to male children, thereby gaining the apellation, umm al-walad (mother of the son), a term not used for free wives and mothers.

The difference in the vocabulary used for men and women indicated their disparate origins. While men came from lands further afield, women generally came from the Christian territories adjacent to al-Andalus. However, the specific origins of male ‘Franks’ and ‘Slavs’ are almost impossible to ascertain. According to Golden, ‘Saqlabī’ was the ‘designation in Medieval Arabic sources for the Slavs and other fair-haired, ruddy complexioned peoples of northern Europe’ who could therefore originate anywhere from the domains of the Carolingians to the upper reaches of the Volga.Footnote ⁵⁴ They were frequently confused with the Rūs, the Majūs, and the Franks, and their deracination, once they arrived in Muslim territory, makes any analysis of their origins impossible. On the rare occasions where evidence exists, it can be surprising. For instance, it is likely that Mujāhid, the ‘Slavic’ ruler of Denia, who emerged from the ranks of the Umayyad and ^cAmirid Saqāliba in the eleventh century, was actually of Italian origin.Footnote ⁵⁵

Debates about the geographical origins and religious identity of the Umayyad ‘Slavs’ also reflect the fact that various commercial routes existed, with sources in different parts of northern Europe. Golden mentions that Slavs were sold in the Crimea, Khazaria and Volga Bulgharia, whence they were transported to the Maghrib and al-Andalus, and Bosworth adds that the Slavs purchased by the Fatimids in the tenth century came from the slave markets of Prague, via Dalmatian ports.Footnote ⁵⁶ Jean de Gorze’s biography makes reference to merchants of Verdun who knew ‘Spain’, and to the commercial route which ran down from Verdun to the Rhone valley and Marseille, from where merchants took ship to Barcelona and the first Muslim port in al-Andalus, Tortosa.Footnote ⁵⁷ Barceló gives this as another route for the trade in Saqāliba, adding that some slaves were taken to Cordoba, but others were re-exported to the east via Almeria, a more southerly port than Tortosa.Footnote ⁵⁸ The existence of a Frankish-Andalusi axis is corroborated by Guichard and Meouak who also mention the various merchants involved: Jews from Septimania and al-Andalus, and ‘Greeks’ of Naples, Amalfi and Venice.Footnote ⁵⁹

Military and elite male slaves begin to appear in the accounts of the reigns of early-ninth-century amirs. ^cAbd al-Ramān II inherited a small ‘Frankish’ mamlūk guard from his father, al-akam I, and this formed the basis for his own larger bodyguard.Footnote ⁶⁰ These soldiers were collectively known as the ‘Silent Ones’ (khurs) because they did not speak Arabic, a title which encapsulates their complete separation from their past and their lack of a voice in their host society. The high point for the recruitment of Saqāliba was, however, the Umayyad caliphal era (929–1031) when their number may have reached 14,000. Numerous high court officials and commanders of Saqlabī origin appear in the historical record, including the courtiers Durrī and Nar, and ^cAbd al-Ramān III’s foremost military commander, Ghālib. However these individuals, like their khurs predecessors, operated within an Arabo-Muslim nexus, and there is no extant evidence that their ethnic and religious origins had any impact upon Andalusi society or culture. That being said, the Risāla of Ibn García, a Hispano-Muslim writing under the patronage of the Saqlabī ruler of Denia shortly after the fall of the Umayyad caliphate, makes the case for the contribution of non-Arabs to Arabo-Islamic culture, using the stylistic form developed by the Persians, who had made a similar case in the early ^cAbbasid era.Footnote ⁶¹

Women from the north were able to have more direct impact upon Muslim Andalusi society. The majority of women mentioned in the sources came from northern Iberia, the Pyrenees and Navarre. They were captured as a result of cross-border raiding, or sometimes presented as hostages for the good behaviour of their men folk. They entered elite households as wives, concubines and domestic servants, and came to play an important role in elite domestic society. Those who produced children often spoke to them in their native languages, creating concerns, among the Arabized Muslim elite, that Romance was ousting Arabic at the highest echelons of society.Footnote ⁶² This affected the Umayyad family itself, since the majority of the mothers of amirs and caliphs were of northern origin. The twelfth- century historian al-Marrākushī, writing about the time of Ibn Abī AÁmir al-Manūr reported that, due to repeated campaigns against the Christian north in the late tenth century, al-Andalus was literally flooded with girls of northern Christian origin. Fathers of Muslim girls could only marry them off by offering highly inflated dowries to counter the appeal and cheapness of these Christian girls to Muslim men.Footnote ⁶³

Conclusion

The inhabitants of al-Andalus between the early eighth and early eleventh centuries viewed the non-Muslim inhabitants of more northerly climes according to the general Muslim worldview of the time, and their own particular relations with such peoples. Two main registers can be identified in the Arabic historical record, and indeed in the analogous Latin one. First, a bi-polar understanding of the world as divided into the dār al-islām and the dār al-harb, which made northern lands areas of ‘unbelief’, ‘polytheism’, and religious wavering, and their inhabitants a primordial enemy. This was, however, an extremely formulaic understanding of the relationship, which was adopted to discuss military campaigns in general, and the annual campaigns undertaken by the Umayyad rulers in particular. In other contexts, the Muslims of al-Andalus appear to have viewed the north simply as a geo-political space ruled by other regimes, such as the Frankish Carolingians and Germanic Ottonians. This space was less developed than the Mediterranean world, as a result of its location in a more northerly and thus less favourable clime. The peoples of the north were certainly perceived to be less civilized than their Muslim and Mediterranean Christian counterparts, enabling Andalusis to place them in the same category as other enslavable peoples such as Turks and sub-Saharan Africans.

Andalusi interest in the world to the north of their frontiers was certainly limited, and initial contacts reflected political expediency rather than the spirit of enquiry. However, the Viking raids of the ninth century sent a frisson of interest through society, and later contributed to the emergence of al-Ghazāl’s fictive encounter with the Vikings, which indicated some rudimentary knowledge of the northern fringe of Europe. The tenth century witnessed increased diplomatic contacts with the Carolingians and the Ottonians, which generated knowledge of both a practical and academic nature, linked to the imperial aspirations of the Umayyads during this era. At the same time, the arrival of increased numbers of enslaved Saqāliba, and the extensive use of Christian concubines, heightened the impact of northerners within Andalusi society itself.

Although not always recognized in the historical and literary record, northerners thus played diverse roles on the Andalusi stage. At one extreme, Christian women from the peninsula were an integral part of many elite families. At the other extreme, the Vikings lived in the Andalusi literary imagination as an archetypal other. In the middle stood the soldiers, merchants and diplomats who travelled between the two worlds, finding much in them that was common. As Daniel states in his analysis of Jean de Gorze’s experience in Cordoba, what is noteworthy is that he did not find it more strange.