Thomas F. Glick



1. Ethnic Cleavages and Cultural Boundaries

[165] Much of the drama and vigor that typified social relations in medieval Spain resulted from the fact that the peninsula possessed a degree of ethnic diversity unknown elsewhere in Europe. This drama was a function, in part, of the degree of cultural difference separating groups, and, in part, of the social, economic, and demographic force of each of the peoples involved. In al-Andalus, the Muslim majority was composed of three powerful groups: the Arabs, a numerically slight but powerful, dominent elite; the Berbers, outnumbering the Arabs, and powerful militarily; and the Neo-Muslims, muwalladûn, eventually the majority of the population. Political life in Islamic Spain throughout the high middle ages was conditioned by the balance of power among these three culturally distinctive groups. In addition, there were two very large non-Muslim religious minorities, the Christians (Mozarabs) and Jews, who played social and economic roles that were not insignificant. In Christian Spain, from the late eleventh century on, conquest entailed the ingestion of large Muslim and Jewish enclaves whose role in society, by virtue of numbers and economic function, was significant.

In general, the historiography of intercultural relations in medieval Spain suffers from two basic misapprehensions. The first is that ethnic conflict and cultural diffusion are mutually exclusive phenomena. Accordingly, those historians who tend toward a conflictive view of the Spanish middle ages either play down the positive features of culture contact or are forced to look for lulls in the storm, necessary in this view for cultural borrowing to occur. The second fallacy is to equate acculturation (a cultural process) with assimilation (a social one) and to assume that the lessening of cultural distance must perforce result in the diminution of social distance -- to confuse ethnic with social cleavages without realizing that these can be cross-cutting. From this fallacy follow a large number of generalizations concerning the cultural fusion of ethnic groups in alAndalus which political facts seem to contradict. In any case, in both Islamic and Christian societies relations among ethnic groups and [166] between members of dominant and minority religions were sharply structured, according to both formal and informal rules and conventions. It is the aim of this chapter to identify patterns of intergroup relations based primarily on ethnic cleavages and to understand how these related to social stratification in the society at large.

Medieval people, whether Christians, Muslims, or Jews, perceived ethnicity largely in terms of religious affiliation. In religiously heterogeneous societies, such as al-Andalus, or Castile (after 1085), religious minorities were perceived as ethnically differentiated units, no matter what the level of their general acculturation to dominant norms may have been. Looking at the broader picture of Christian-Muslim confrontation, Christians perceived an absolute dichotomy, of an ethnic nature, between the moro (that is, the Andalusi Muslim, whether Arab, Berber, or Neo-Muslim, white or black) and the cristiano (as an ethnic category, a speaker of Romance from the north of Spain). Among the Muslims, one notes the same foreshortening of ethnic recognition: the Christian enemy was an unbeliever (kâfir) or a polytheist (mushrik), a supra-ethnic category which was a sufficient and universal definition.

Before the advent of modern nationalism, which accentuated the role of language in ethnic ascription, linguistic differences, although certainly noted, were not considered an insurmountable bar to assimilation.(1) The general acceptance of Latin as a lingua franca among all Christians and of Arabic among all Muslims created an atmosphere in which, for example, Romance dialectalisms were accepted as variants all more or less equidistant from the norm. In the ethnically plural societies of the peninsula, linguistic differences were accepted as part of the cultural landscape and accommodations easily made to minimize their role as barriers to communication, whether in politics or in the marketplace (for example, the generalized bilingualism of merchants; specialized officials who served as translators, scribes, and diplomats). On the other hand, languages inevitably picked up connotations arising from ethnic conflict -- for example, the use of the term cristiano in the sense of "Romance language," that spoken by Christians.

Figure 2 illustrates some of the nuances of ethnic ascription in medieval Spain, with regard to subordinate enclaves in each society. The primary reference point was religion, and the crucial differentiating factor was whether a group, after conquest, had remained faithful to its original religion or had converted. A second point of differentiation, more[167] important to Christians than to Muslims, was whether a person spoke a language different from that typically associated with his religious group. These terms, first by their very variety, indicate that there was a considerable range of socially recognized statuses along the roads to acculturation and assimilation, and, second, by their specificity, that such groups participated in a structured, patterned system of group relations. Another family of terms, not included in the figure, related to those people who fitted into no specific one of these groups but who were cultural commuters (or, with the negative connotations attached to such people, renegades) who passed back and forth between Islamic and Christian societies and had no particular allegiance to one or the other. That such people (called enaciados or elches -- the latter from Arabic 'ilj, "renegade") existed at all was grating to the sensibilities of the times, which required that groups be plainly identifiable, by virtue of language, religion, tribal affiliation, and so forth.(2) There follows from this mentality, amply documented below, that conversion into one's group was, theoretically at least, encouraged, while conversion out was forbidden and liable to harsh punishment [168]

2. "Protected" Religious Minorities

The Koran -- a syncretistic book with elements drawn from Judaism and Christianity -- envisioned a polity which was stratified along religious lines. This was an accurate representation of Arabian society in the time of Muhammad, when pagan Arab tribes coexisted with Christian and Jewish ones. Thus the Koran describes a society dominated by Muslims (assumed to be Arabs), with provision made for approved religious minorities, characterized as dhimmis ("protected" peoples) or "People of the Book" (those with a revealed scripture recognized by Muslims as divinely inspired -- that is, Christians and Jews). When the Arabs burst forth from the Arabian peninsula and extended control over great numbers of non-Arab, non-Muslim peoples, they found themselves obliged to encompass this vast religious and ethnic diversity within the Koranic framework, as further defined and elaborated in the growing corpus of Islamic law.

The meticulous legal structuring of social and economic relationships between the dominant Muslims and the subordinate Christians and Jews who found themselves within the Empire underlies the generally harmonious interaction of religious groups in the Islamic middle ages. Dhimmis suffered civil and legal disabilities in comparison with Muslims, but they also enjoyed the security that inhered in formal juridical status. Three ramifications of this status are particularly relevant to the situation of Christians and Jews in al-Andalus. First, by accepting status as dhimmis, members of religious minorities were effectively excluded from participation in political power. Although, under special and always temporary conditions, Christians and Jews did attain considerable political power, the only way to achieve substantial upward mobility in the society at large was to convert to Islam. Second, the system had built-in inducements to assimilation in the form of tax incentives (exemption from the jizya or poll tax) for those who converted. Third, the formal recognition of religious differences meant that Islamic society sanctioned among religious minorities a relatively high degree of cultural enclosure. That is, by granting the persistence of the primary factor that made peoples distinct from one another --religion -- Islamic society put less pressure on religious minorities to accept its values than it placed upon non-Arab Muslims.

When the Christians came to determine what their own comportment would be towards religious minorities (a modest number of Jews before 1085, large-and increasing-numbers of Muslims and Jews thereafter) [169] they used the ready-made example afforded by Muslim treatment of dhimmis. As in al-Andalus, the religious minorities were accorded legal safeguards which promised the maintenance of a relatively high degree of group autonomy (again, connoting the acceptance of strong cultural boundaries), reinforced by spatial segregation in juderías and morerías and a detailed law code defining intergroup relations. This system, in Christian guise, worked substantially less to the favor of the minorities than it had in its original form. The reason for the discrepancy, and the erosion of the social status of subject Jews and Muslims from the start, lies in the differing concepts of law prevailing in the two cultures. Islamic law, although subject to a relatively limited range of differing interpretations, was universal and unchangeable. The safeguards afforded to the People of the Book were the norm, and suffered erosion only at the hands of weak rulers, who permitted unlawful contraventions of minority rights, or of despots whose fanaticism led to direct contraventions of law.(3)

In the Christian kingdoms there was no general norm, except for vague guarantees of freedom of worship and group autonomy, but rather the rules were pacted with local groups by specific rulers and were subject either to ratification or change by each successive ruler. This meant that the actual situation of Jewish and Muslim minorities differed widely from place to place and was, moreover, highly sensitive to change in response to social and economic interest groups who were able to gain the ear of the monarch. The insecurity and precariousness inherent in such a situation accounted in great part for the steady erosion of group cohesion among religious minorities under Christian control.


The intent of the dhimma arrangement and its imperfect Christian counterparts was to ensure that religious groups were kept separate, distinct, and apart from one another, lest the dominant religion suffer contamination from the subordinate ones. Here we can only cite a number of arrangements that were standard in both Islamic and Christian societies. The keystone of group autonomy was that each religious community would be ruled by its own body of religious law, in its own courts, by its own judges, chosen by the community itself. The jurisdiction of the autonomous communal courts was limited to civil cases involving members of the community both as plaintiff and defendant; those which involved a [170] member of the dominant religion generally had to be tried by a state judge, as did most criminal proceedings. Thus the religious courts handled cases involving all aspects of family law (marriage, divorce, inheritance) as well as other social and economic litigation falling under the purview of religious law.(4) The autonomous religious court system had obvious cultural and social repercussions: the court was a fulcrum of group cohesion and served to reinforce the distinctive cultural traditions of the group. Thus, in Jewish courts in al-Andalus, Talmudic law interpreted according to the Babylonian school served as a mechanism to keep alive a particular form of Judaism and there, as throughout the Islamic world, communal authorities condemned (and even punished) recourse to Islamic courts.(5) The Mozarabs also had an autonomous court system, with a judge (censor) who dispensed justice according to the Forum Judicum as well as, ostensibly, canon law.(6)

In Christian Spain, autonomous courts for Jews and Muslims were provided, but the tendency was to undermine, whenever possible, the prerogatives of the religious minorities in cases involving a Christian. Thus Alfonso VII decreed in 1109 that all suits between Christians and Jews in Toledo would be heard by a Christian judge, even when a Jew was defendant, a trend that was continued until in 1286 the Cortes of Palencia petitioned the king to deny the Jews any special judge at all, but should have an ordinary (that is, Christian) judge hear their pleas separately, so that Christians and Jews alike would have their due.(7) Likewise, with autonomous Islamic courts, there was a range of solutions for the resolution of suits between Muslims and Christians, from autonomous judge, to special judge, to only a Christian judge (in Toledo).(8) In thirteenth-century Valencia the formal structure of Islamic jurisprudence was maintained relatively intact, with qâdî, amîn, and sâhibal-madîna (the latter two were enforcement officers), and King Peter was explicit in his desire that "Saracens ought to fall under the jurisdiction of their qadi" (1284).(9)

The Christian law codes, whatever else their failings, seem to have been especially punctilious in defining the manner in which the correct oath should be given by people of the three religions. The Fuero of Calatayud specifies that in inter-ethnic cases a Christian must swear on the cross; the Jew, holding the Torah; and the Muslim, by repeating the words "I swear faithfully three times." The Fuero of Jaca specifies that in cases involving people of different religions, each must defend himself according to his own law, even in cases of assault, and, further, that agreements between [171] people of different religions ought to be drawn up by the scribe of one religion or another, depending on the kind of issue.(10) James I, who played a role in the revival of Roman law in the Kingdom of Aragón, also had clear notions of how Semitic law ought to function; he regarded the sharî'a (Islamic law in general) and the sunna (custom of the Prophet, but here understood as customary law in general, as opposed to canon or civil law) as still in force among the religious minorities. In the charter of Denia (1227) James defined sunna as the "privileges and customs which the Saracens were accustomed to have in the Kingdom of Valencia." He understood Jewish law in the same way, once referring to the "sunna of the Jews" (a characterization that reflects the strong Arabization of Spanish Jewish culture in the thirteenth century as well as the monarch's own confusion). In spite of these tolerant expressions, by the end of the century the system was beginning to show substantial erosion as basic principles such as the right to have witnesses of one's own religion were increasingly abandoned.(11)

Although the court system and maintenance of religious law were the cornerstones of minority autonomy, there were other legally enforced social mechanisms designed to minimize intimate contact among peoples of different religions. Intermarriage was prohibited by both Christian and Jewish religious laws. The fact that Muslim men were allowed to marry infidel women contributed to the numerical attrition and sense of threat within the minority enclaves and heightened their own internal strictures against intermarriage. Castilian law codes stipulated harsh penalties for fornication between Christian women and Jewish or Muslim men (leaving unmentioned the license of Christian men to exploit minority women sexually).(12)

Dominant religious groups naturally provided harsh penalties for apostasy. The Siete Partidas provides the death penalty for the Christian who turns Jew or for him who, having lost his mind and desiring to live in the Moorish style, becomes a Muslim.(13) Generally, the aversion to conversion to any but the dominant religion was extended to lateral conversion among the religious enclaves: for example, Islamic law forbade the conversion of Christians to Judaism (and vice versa), although Jews in fact continued to convert their slaves. The same stricture is found in the Partidas, where Jews are forbidden to convert their slaves or captives even when these were "Moors or other barbarian peoples." This stricture was in addition to the general provision that neither Jews nor Muslims should [172] proselytize Christians but should keep their own faith, without abusing that of the majority. Linked to the ban on proselytization were provisions, typical in all Castilian law codes, forbidding Christians to dwell in the same house with Muslims or Jews.(14)

The question of how much power a member of a religious minority might have over people of the dominant religion was subject to a range of legislative prescriptions, ranging from prohibition of holding a slave of the dominant religion, to the holding of public office, and the more complicated matter of economic relations, particularly lending money at interest. It was standard that no minority person be allowed to own a slave belonging to the dominant religion.(15) As for public offices, the general pattern of the appointment of minority men to high rank, particularly in financial bureaus, through the personal patronage of a ruler and the objections to such procedures by religious leaders, on religious grounds, and by those of the dominant group who were competing for power with the favored officeholder, or those of the dominant religion who complained of the humiliation suffered at the hands of such favorites, was common to both Islamic and Christian society. In the early Emirate, Mozarabs held positions of power: al-Rabî', the head of the Christian community of Córdoba, served al-Hakam I. As the power of fuqahâ', the Muslim jurisprudential class, grew, that of the Mozarabs diminished. In the reign of Muhammad I, Christians were dismissed from the militia and their role as state officials dwindled. Jewish aristocrats, such as Hasdai ibn Shaprût, served the Umayyads in a variety of roles, typically as ambassadors and physicians.(16)

A special case of Jewish preferment arose in eleventh-century Granada, when two members of the ibn Nagrella family, Samuel and Joseph, served the Berber Zîrîd rulers as vizier. The power achieved by these two men was based on a precarious balance of power between the Berber rulers and the Arab elite in the kingdom. The Jews represented the economic strength of the middle class of Granada and rose to high office (the highest ever held by Jews in Islamic Spain) as a counterbalance to the Arab aristocracy. The Zîrîds, then, were able to maintain power by merging a Jewish-dominated fiscal and administrative apparatus with a strong Berber militia.(17)

In Christian Spain a similar pattern was evident. Jews served individual rulers, particularly as tax officials, in spite of protests by religious zealots and by those to whose interest Jewish domination was perceived as [173] threatening. Alfonso VI of Castile's preference for Jewish civil servants brought him a warning from the Pope in 1081 not to grant Jews positions of authority over Christians. Nevertheless, he appointed Joseph Ferrizuel ("Clidellus") and his nephew Solomon to a number of high positions later in the same decade. In Aragón, Yahya ben David of Monzón served Alfonso II (1162-1196) as a bailiff, and Sheshet b. Solomon ("Perfet") had served count Ramon Berenguer III (ca. 1121) in the same office. Such preferment was continually attacked, to no avail, especially in Castile, where various law codes reiterated the Visigothic prescriptions (of the Third and Fourth Toletan Councils) forbidding Jews to hold public office.(18)

Jewish power was increasingly related to involvement in financial transactions, particularly the lending of money. Medieval custom and religious law took a dim view of lending money at usury to a correligionary. Jews could not take interest from other Jews, but they could from Christians or Muslims. The same, of course, was true of Christians, whose habitual recourse to Jewish capital needs no comment. Lending does not seem to have caused as many problems for Jews in the money economy of the Islamic world as it did in Christian Spain. Although money-lending did not become a pronouncedly Jewish activity in a place like Castile until the fourteenth century -- before which time Jews were engaged in a multiplicity of other economic pursuits as well -- the mention of two Jewish money-lenders from Burgos in the Poema de mio Çid confirms that this stereotype was firmly set by as early as the mid-twelfth century. The tolerated interest for Jewish loans, set by statute repeatedly in the second half of the thirteenth century, was thirty-three and one-third percent, which led early in the next century to repeated petitions in the Cortes for relief from debts owed to Jews.(19)

The involvement of Jews not only in money-lending but in a host of middle-class commercial and industrial activities made this group particularly susceptible to popular hostility in times of economic recession, leading Antonio Ubieto to use incidences of anti-Jewish (or anti-minority) legislation and policies as a rough indicator of periods of economic decline. Thus Ubieto sees as symbolic of recession the persecution of Mozarabs during the reign of 'Abd al-Rahmân II; the persecutions of Christians and Jews in Córdoba (1013), Zaragoza (1039), and Granada (1066); the expulsion of Mozarabs by the Almoravids, concurrent with the currency devaluation of the 1120's; and anti-Jewish outbreaks in [174] Córdoba (1135) and Valencia (1144-1145). In Christian Spain, the depressions of the twelfth century brought increasing restrictions on Jewish activities and the general persecution of 1275 and anti-Jewish dispositions of the Cortes of 1301 and 1307 can also be related to economic recessions.(20)

There are also general indices of the low esteem of religious minorities, such as the setting of the Jewish wergild, formerly at the level of a minor noble, at parity with that of a peasant (by Alfonso VII in 1109), as well as differential standards for punishment: in some Fueros murder was a capital crime for Jews and Muslims, but subject to only a fine for Christians.(21)

These examples, selected from among many, well illustrate the systemic aspects of minority relations in both Islamic and Christian Spain. According to prevailing norms, minority religions were kept at a distance from majoritarian society. The communal autonomy of these groups, often represented as the very symbol of tolerance, was in fact the institutional expression of ethnocentric norms which held such groups in abhorrence, as tolerated but alien citizens who were not to share in social life on the same basis as members of the dominant religion. It is clear that such institutionalized forms of enmity would not protect the minorities in the case of severe economic or social disruptions when, as perpetual outsiders, they easily became scapegoats. The point is made well by Robert I Burns, who remarks, apropos of Muslim minorities in Valencia, that "the tolerance that fully preserves an alien group's existence also condemns that group to a life apart, inverted, and undernourished, so strange to the other community as to seem repugnant and even inimical."(22)

Further systemic ramifications of the semi-isolation of religious minorities are apparent in the effects of social segregation upon their culture. Communal autonomy and social segregation influenced the culture of the minority groups in two contradictory ways. The guarantees of group autonomy and the external pressures placed upon the group to maintain a relatively high level of group identity for purposes of control had definite solidarity-inducing effects upon the minority. On the other hand, second-class citizenship and the social, economic, and in some cases cultural, prestige of the dominant religion induced pressures for acculturation and for the abandonment of traditional norms. These ambivalent influences are well illustrated by comparing the cultural contours of the Jewish and Christian minorities in al-Andalus.

We can agree with S. D. Goitein that for Jews the possibilities for [175] acculturation were greater in the Islamic than in the Christian world because the high culture of the former was to a great degree secular in nature. Throughout the Islamic world, Jews had become completely Arabized in culture and in language by A.D. 1000. The result was that, in spite of legal restrictions already mentioned, not only did Jews participate fully in Arabic culture, but their own distinctively Jewish traits were invigorated and enhanced by creative interchange. Thus, for example, both Hebrew secular poetry and philology were substantially revitalized by Andalusi Jews under the influence of and stimulus of Arabic poetry and grammar.(23) Moreover, much of Jewish philosophical and scientific scholarship was written in Judeo-Arabic (Arabic written in Hebrew characters), which developed into a rich vehicle for literary and scholarly writing, since it was not constrained, as normative classical Arabic was, by the model of Koranic language, which imposed conventional limitations upon Muslim prose writing. The social status of scholarship was also influenced by the Islamic milieu, in which both scholarly and commercial pursuits conferred prestige and were frequently combined in the same persons. The figure of the "learned merchant" as bearer and seeker of high culture on his far-ranging business trips was as characteristic of the Arabized Jew as of his Muslim counterpart.(24)

Yet in spite of the great vigor of the cultural and economic life of Jews in al-Andalus and the infrequency of violent outbursts against them, the psychological burden of second-class citizenship weighed upon them as it did upon the Mozarabs. An Andalusi Jewish merchant writing around 1140 from Fez to his father in Almería (a hub of Jewish economic activity and also a place where, possibly as a consequence, anti-Jewish feeling ran high) that hatred of Jews "is rampant in this country to a degree that, in comparison with it, Almería is a place of salvation."(25) The subsequent Berber invasions had predictable consequences for the Jewish community.

When, as a result of those invasions (eleventh and twelfth centuries), the vast majority of the Andalusi Jews migrated to Christian Spain, their cultural position changed abruptly. Their secular culture was incongruent with the religiously oriented high culture of the Christians, as a result of which the disorientation of Jewish intellectuals was inevitable. A poignant example is Moses ibn 'Ezrâ, poet and courtier of Granada, who arrived in Castile in 1095, wandering from town to town. "I have come," he complained, "to the iniquitous domain of a people scorned by God and accursed by man, amongst savages who love corruption. . . " Nurtured in [176] his youth in "the gardens of truth" (the Arabic culture of Granada), he was now obliged to "hew, in old age, the wood of forests and folly."(26)

Others were able to make the transition on the basis of their mastery of Arabic culture. (The role of men of science will be discussed in a later section.) But those who served the Christian kings and princes (in this case, Sheshet Perfet) were described by Judah ibn Tibbon as having attained "fame and high rank only through the writing of Arabic."(27)

Thus the very men who had felt, through their domination of Arabic cultural forms, part of Andalusi society had, when transferred to Christian lands, as a function of that very learning been reduced to the status of cultural go-betweens -- valuable for that to the Christians but condemned to ambivalence and insecurity in a cultural climate for which they had no feeling and in which they had, as yet, no creative role to play in and for themselves.

The Mozarabs, those Hispano-Romans who had not converted to Islam but who had remained in al-Andalus, formed substantial minorities in Córdoba, Toledo, and other large cities, as well as in many areas of the countryside. The extent of their Arabization is the subject of debate among historians. The centerpiece in the argument for their substantial Arabization is a much-cited passage from the Indiculus luminosus of Paul Albar, which states that the Christian youths of ninth-century Córdoba were "highly regarded for their ability to speak Arabic," were conversant with Arabic literature, did not know their own law, nor could they speak their own tongue (Latin).(28) Yet, Albar's testimony is almost unique. Mozarabs created no Christian literature in the Arabic language (or at least none has survived), and their rather extensive Latin literature reveals a minimal interest in the knowledge of the Islamic society around them. An apt example of the cultural isolation of the Latin Mozarab writers is that their knowledge of the life of Muhammad came from a debased, distorted Latin version brought back from the monastery of Leyre in Navarre by Eulogius, who could have obtained more accurate information by asking any Muslim in the street. Based on an analysis of this literature, together with the fact that the term Mozarab (from Arabic musta'rib = "Arabized") first appears in Latin texts and seems not to have been used by Muslims, Vicente Cantarino argues that Mozarab was initially a pejorative term used by the fanatical leadership of the martyrdom movement to denigrate more highly acculturated Christians. In this view, the leadership purposely cultivated ignorance of Arabic-culture and, [177] because of its inability to interact creatively with Arabic culture, condemned the entire group to "historical irrelevance."(29)

Cantarino's portrayal of a group actively resisting acculturation is undoubtedly true of a substantial segment of the elite. Other aspects of Mozarabic society, such as its receptivity to heterodox religious trends, also contribute to a picture of a culturally disoriented group. Yet, this cannot be the entire picture if only because the demands of the marketplace alone would have been a sufficient enough inducement for those Mozarabs who produced marketable goods (craftsmen and farmers) to have learned Arabic. In this respect evidence pertaining to Mozarab immigrants to León in the ninth and tenth centuries is relevant. The documents reveal a typical pattern, extending into the eleventh century, of men with Arabic names married to women with Romance names. Of fourteen mixed-name couples in a collection of charters from the Monastery of Ardón between 947 and 989, all but one manifest the pattern alluded to (e.g., Aiza and Argentea; Yahea and Filoria; Abuhab and Vistrildi; Zuleiman and Loba, etc.).(30)

This pattern is what one would expect in a multicultural situation such as that prevailing in the areas of al-Andalus whence many of these settlers originated: men who were bilingual Romance and Arab speakers, women who were monolingual Romance speakers.(31)

Lacking the congruence with Islamic culture that enabled the Jews to acculturate rapidly, the Mozarabs adopted the language at a slower pace; but by the time Toledo was conquered in 1085 the indigenous Christian population was composed wholly of monolingual Arabic-speakers. Those who migrated to León earlier, needless to say, had no difficulty adapting to the prevailing cultural norms. Their role as bearers of elements of Islamic culture will be discussed in due course.

The cultural situation of the Mudéjares, the Muslim minorities in Christian Spain, is somewhat akin to that of the Mozarabs. The Mudéjares of the twelfth and thirteenth centuries appear uniformly unwilling to adapt to Christian culture. They attempted to maintain their language and traditional culture intact, in the face of ever-increasing pressures towards acculturation from the Christian authorities. In the period 1085-1300, these pressures did not reach the level or have the impact achieved in later times, but in subtle ways the fabric of Muslim society under Christian domination was slowly altered, as traditional Islamic institutions -- whose survival was guaranteed under the terms of capitulation agreements -- were increasingly colored by, or impinged upon, by Christian practice. Thus village [178] elders, who played a distinct role in a tribally based social structure, were viewed by the Christian authorities as having the attributes of a town council, although there was no precedent for such a body in Islamic law. The Christians thus forced a new role upon the Muslim communities and institutionalized it.(32) The Mudéjares, perhaps because they were viewed as a fifth column, were subject as no other peninsular religious minority to formal pressures for cultural change, directed by the rulers -- a strategy which in the Islamic world was irrelevant to the dhimma contract and where the motivation for change originated largely within the minority communities themselves.

3. Ethnic Competition in al-Andalus

Although Islamic law made clear provision for the protected religious minorities, it did not spell out in any systematic or definitive way the status of ethnic groups who accepted Islam but resisted acculturation to Arabic norms. Anthropologists have noted that egalitarian creeds have generally proved ineffective in preventing ethnic stratification, and the Islamic experience bears out this conclusion. Although Islamic law envisioned a universal brotherhood of believers, in point of fact ethnically stratified social systems evolved in many regions of the Islamic world. The usual method of dealing with the assimilation of ethnic minorities was to accord the neophyte the status of client (mawla), whereby new converts (often entire tribal groups) would attach themselves to a powerful Arab family or tribe, adopting its lineage and social status.

In ethnically plural societies, access to power usually involves leaving behind the culture of one's original reference group and adopting that of the dominant group. But because they were Muslims, the ethnic minorities of the medieval Islamic world already had putative equality in the eyes of the law and, thereby, legal guarantees of access to power on a par with other believers. Therefore, in such areas as Persia, North Africa, and Islamic Spain, where Arabs were outnumbered by indigenous peoples of markedly different culture, political life was typified by competition for power among Muslim groups. One factor controlling success in competition was the ability of non-Arabs to master certain cultural and social skills associated with Arabs. As a result, cultural boundaries of ethnically differentiated, competing groups of Muslims tended to be more blurred than among non-Muslim groups, inasmuch as a certain degree of assimilation was prerequisite to success in inter-group competition.

[179] Thus in Islamic Spain, in addition to differentiation among groups along religious lines, with ethnic ramifications, the Muslims themselves were split into three groups, all with putative access to power according to the dictates of Islamic law: Arabs, Berbers, and Neo-Muslims. The demographic structure of the Muslim, sector of Andalusi society virtually mandated intense competition among its components. The Arabs were a ruling, bureaucratic, and landholding elite, but were scant in numbers: Târiq's original invasion force of 12,000 men was said to include but seven.(33) Subsequent emigration under Umayyad impetus was mainly felt at the very top of society and could not have increased the Arabs' demographic weight significantly. More numerous were the Berbers who joined the Arab command in North Africa and constituted the bulk of the invasion force. Despite the withdrawal of substantial numbers during the drought and famine of the 750's, fresh Berber migration from North Africa was a constant feature of Andalusi history, increasing in tempo in the tenth century. Hispano-Romans who converted to Islam, numbering six or seven millions, comprised the majority of the population and also occupied the lowest rungs on the social ladder. Yet, through the system of clientage, substantial upward mobility was possible, especially among the Neo-Muslim elite.

The political history of Islamic Spain from 711 until the collapse of the Caliphate has been typically portrayed as a steady and logical evolution of a new state from political infancy as a province of the Caliphate of Damascus, through an adolescent period as a fledgling emirate, to the mature and fully developed Caliphate, a culminating point whose fulfillment was interrupted and stymied by numerous revolts by dissident factions usually Berbers or Neo-Muslims, or coalitions of these with disaffected Arab groups. Emphasis on the attainment or obstruction of political stability tends to obscure the dynamics of the competitive ethnic system that the state sought to control. The failure of the Umayyad state can then be presented as the tumbling of a house of cards, a building whose structure was flawed, the rupture (in one historian's characterization) of the "fictitious equilibria" upon which the Caliphate was built.(34) This, indeed, is the popular wisdom. Ibn Khaldûn asserted at length that "a dynasty rarely establishes itself firmly in lands with many different tribes and groups."(35) Modern sociological intuition is more nearly the opposite: many centralizing states rule over a diversity of peoples and their success in ruling is greater when there are more rather than less cleavages. The fewer the number of powerful groups, the greater is the [180] possibility for conflict between them. Therefore, we shall argue that the equilibria among ethnic groups upon which the Umayyad state were based were unstable (although by no means unstructured) and not, for that, fictitious. Their reality accounts both for the political fortunes of the Caliphate and its ultimate demise. (See Chapter 6, sections 1 and 2.)

The line of differentiation between Arabs and Berbers is admittedly an indistinct one; or, at least, it became so in short order after the Arab conquest of North Africa, as certain Berber elements were quick to assoclate their destinies with that of the elite. The most visible cleavage was linguistic, involving distinct and unrelated language groups. The process of "Arabization" which began at this time involved the learning of the Arabic language and was much slower in pace than the associated process of religious conversion -- Islamization. The process was largely, but not completely, one-way, the rate of its progress varying according to political and social factors. For a moment in the eleventh century the direction of cultural flow was actually reversed, a case in point being the Hammûdid Arabs, pretenders to the Caliphate allied to the Zanâta Berbers, who were Berber-speakers and spoke Arabic with a Berber accent.

The fact that there are few Berberisms in Castilian and most Berber-derived place-names are derived from tribal appellations (including some in Old Castile, where Christian Berbers appear to have settled in the eighth century) has led to the general conclusion that there were no purely Berber-speaking centers in al-Andalus by the tenth century.(36) But such a conclusion is problematical. It is more likely that there was extensive bilingualism and biculturalism among Andalusi Berbers. Instead, there must have been a range of acculturation. Those (relatively few) who lived in cities, the literate class, may have been totally Arabic-speaking. On the other hand, the Arabization of rural, isolated Berber nuclei in such mountainous backwaters as Teruel and Albarracií is improbable. So also, in those of the Party Kingdoms ruled by Arabized Berbers of long residence in the land (the Aftasids of Badajoz, the Dhu'l-Nûnids of Toledo) it seems probable that only the aristocracy and middle classes were fully Arabized, while the masses remained monolingually or bilingually Berber-speaking. Even if we assume substantial Arabization, there is ample evidence that acculturation did not bring with it any general lessening of social distance between Arabs and Berbers. We have already commented on the economic division of labor between the two peoples, and this was reinforced by social and religious differences.

[181] The ascription to the Berbers of religious heterodoxy was a key factor in the maintenance of social distance. Whether or not it was true, as ibn Khaldûn alleged, that the Berbers were guilty of repeated apostasy, such was the image generally held by urban Arabs. Moreover, certain forms of Berber heterodoxy, such as Khârijism, were linked to the issue of discrimination. The Khârijites insisted that the universalist ideology of Islam be put into practice, and they affirmed the non-Arabs could aspire even to the Caliphate. The Arab characterization of Berber religion as a corrupt, unorthodox, and heretical form of Islam was one rationale for the maintenance of social distance.

Cultural, linguistic, religious, and socio-economic cleavages manifested themselves in a series of Berber uprisings, conducted more or less against the central authority, which punctuated the history of the Umayyad state. These included the rising of Shaqya b. 'Abd al-Wahîd against 'Abd al-Rahmân I (768-777); of Asbag b. Wansus against al-Hakam I (805-813); and rebellions of Tawril and other tribes against 'Abd al-Rahmân II (822-852).(37)

Any instance of Berber rebellion, or of political revolt in which Berbers participated as a recognized entity, must be taken as an indication of social distance, as well as of the relative cohesiveness of Berber culture at the time.

The political breakdown which occurred during the reign of 'Abd Allâh (888-912) is significant for the salience of group boundaries. This struggle was initiated in 889 by the powerful Arab chieftain Kurayb b. Khaldûn of Seville (ancestor of the historian) against Neo-Muslim opponents. Both Arabs and Berbers were divided into confederations which, in times of conflict, tended to pair off symmetrically across ethnic lines. Kurayb led a coalition of Yemeni Arabs and their Berber clients in western al-Andalus. These Berbers, led by Junayd b. Wahb al-Qarmûni (of Carmona), belonged to the Branes confederation. By a kind of natural pairing off, this coalition was then opposed by another comprised of Qaysî Arabs, traditional enemies of the Yemenis, Botr Berbers (enemies of the Branes), and Neo-Muslim clients.(38) If Botr-Branes lines were sharply drawn in the closing years of the ninth century, it is difficult to make a case for the virtual fusion of Arabs and Arabized Berbers in the tenth.

A tendency to form political coalitions along ethnic lines was characteristic of the Taifa period of the eleventh century after the polity had split into small, ethnically oriented kingdoms. In 1047-l048, for example, a [182] powerful Berber coalition supported the proclamation of a Hammûdid pretender to the Caliphal throne. This coalition included the rulers of Carmona, Morón, Arcos, Huelva, and Badajoz, and was led by Bâdis b. Habûs, the Zîrîd emir of the Berber kingdom of Granada. All these made common cause and attacked the Arab 'Abbâdids of Seville. The coalition included both Sanhâja and Zanâta elements, in addition to "Arabized" Berbers of long standing such as the Aftasids of Badajoz, a Miknâsa group whose ancestors had participated in the conquest of the country. In spite of their Arabization, however, they nonetheless participated in this Berber coalition.(39) The Birzâlids of Carmona were a Zanâta group of recent immigration, closely allied with the Sanhâja Zîrîds of Granada.

Both the Birzâlids and the Zîrîds had migrated with their families in 997-998 to reinforce al-Mansûr's armies and had settled in the Granada-Jaén area. They were differentiated ecologically, however, the Zanâta occupying the countryside, the Sanhâja settling in the cities. In this situation, interestingly, the Branes-Botr split appears to have been smoothed over and a stable modus vivendi reached which permitted the formation of a common front to oppose the Arab bloc.(40)

Comparing this period with that of the fitna of the ninth century, one notes that in the earlier period, before the massive conversion of Neo-Muslims, a sharp sense of ethnicity is manifest in the perpetuation of traditional moiety groups. In the Taifa period, perpetuation of moiety distinctions in the face of an overwhelming Neo-Muslim majority would have been politically disastrous for Berbers. The completion of the process of conversion, together with the northward migration of substantial numbers of Christians, produced a situation where the polity came increasingly to be split into two monolithic opposing blocs of "Andalusi Arabs" and Berbers, the latter a minority, but with great military power.


The changing relationship between Berbers and Arabs can be appraised through texts which reflect the rising prestige of Berbers in the late tenth century, the result of their growing military and political power. At the same time, it is probable that substantial class differentiation had occurred within the Berber group, involving the emergence of a military elite which sought participation in political and economic power, thus setting the stage for extreme anti-Berber reactions by the lower strata of Andalusi [183] Arabs, fearful of a loss of a status. Anti-Berber feelings were expressed in stereotypes typical of societies in which ethnic groups are competing for power.

A passage from ibn Hayyân sheds light on shifting attitudes towards Berbers in the time of al-Hakam II. Ibn Hayyân stated that 'Abd al-Rahmân III had hated Berbers and that al-Hakam II had begun his reign in the same way, going so far as to prohibit his pages, mercenary troops, and regulars from imitating Berber manners in dress or in saddlery. But then the Caliph, growing. fond of Berber horsemen and recruiting them for his army, departed from the practice established by his predecessor; it was this change in attitude which, according to ibn Hayyân, landed al-Andalus in the morass of civil turmoil accompanying the dissolution of the Caliphate.(41)

The imitation of Berber dress, criticized by Berber-hating chroniclers, is a significant phenomenon which provides a rough index of the changing direction of cultural flow. As the Berbers became more powerful and prestigious, it became more acceptable and common for Arabs to imitate them. According to al-Maqqarî, the religious and intellectual elites in the highland towns of the south and west in the tenth century adopted the Berber headdress, the turban ('imâma), while in Córdoba and the east the oriental high bonnet (kalansuwa) was still the fashion. But, later on, Berber modes were adopted even by the Arab aristocracy of Córdoba. When the 'Amirid 'Abd al-Rahmân Sanchuelo came to power at the end of 1008, Berber prestige was enhanced owing to the influence over the ruler exerted by the Sanhâja chieftain Zawl ibn Zîrî. In January 1009 "Sanchuelo constrained the nobles to present themselves at his palace at [Madîna] al-Zâhira wearing not their traditional high colored bonnets, but Berber-style turbans," a bitter pill indeed for the haughty Cordoban aristocracy. The ultimate in this kind of adulatory emulation was reached when the Almoravid ruler Yûsuf ibn Tashufîn first crossed over into Spain and was entertained at a banquet by al-Mu'tasim, the Tujîbî prince of Almería, who tried to please his guest by sporting a burnus. The Berber-hating poet-king of Seville, al-Mu'tamid satirized this degrading act of subservience in a poem. Recalling his pleasure with some "peerless maiden" the king noted that:

I also recalled-and my bliss was increased
The confusion and shame of the burnus-clad chap!

[184] "Confusion" is a revealing description, suggesting a period of extreme cultural flux, during which competing reference groups struggled for men's allegiance. Two centuries later, with Berber power on the wane, the turban was abandoned by all classes throughout eastern al-Andalus, an indication of a definitive reversal in the direction of cultural flow, and it was worn only rarely by the subjects of the Nasrid kingdom of Granada.(42)

Arab stereotypes of Berbers were typical of situations in which the ruling caste is threatened by a powerful subordinate group. Like all ethnic stereotypes, these reflect -- albeit in an exaggerated and pejorative form -- real social and economic roles. As mountaineers and herders, Berbers were perceived as crude and ignorant. As soldiers and horsemen, they were seen as violent and arrogant. Al-Hamdânî's famous formula whereby violence was apportioned in the ratio of nine-tenths for the Berbers and one-tenth for the rest of mankind is a perfect expression of this aspect of the stereotype.(43)

During the fitna, or period of anarchy at the end of the Caliphate, violent outbursts and ethnic massacres were common, as when the masses of Córdoba, afraid and angered by Berber hegemony in the capital, rose in 1010 and slaughtered Berber troops under Zawi's command, causing the relocation of the Zîrîds in Granada. This type of behavior characterized Arab-Berber relations for the rest of the century. In 1035, the Birzâlids intervened to prevent a massacre of Hammâdids by the Sevillan 'Abbâdids. Similarly, only the intervention of Samuel ibn Nagrella, the Jewish vizier, prevented the Zîrîd ruler from killing all the Arabs in Granada in 1058 in retaliation for the assassination of the Ifranid Berber ruler of Ronda.(44) We have already noted the alliance of Berbers and Jews in the eleventh century. This fact was not lost on Arab poets who joined the Jewish stigma with Berber heterodoxy. "Your sword," wrote a court poet of al-Mu'tadid of Seville, "has raged against a people who have never believed except as Jews, although they are called Berbers."(45)

"Berberization" was a fear very much in the minds of the Arab intelligentsia, who feared both for the bastardization of their cultural traditions and for the supposed threat to orthodoxy. Córdoba during the fitna was described as "a city such that [we pray] that God may forgive its inhabitants' lapse, for they became Berberized (tabarbarû), mingled with the Moroccans, and adopted the creed of the Egyptains," the latter a reference to the Shi'ite inclinations of the Hammâdids.(46)

185] In view of the real threats of Berberization perceived by Andalusi Arabs of the tenth and eleventh centuries, claims for the speedy and thorough Arabization of Berbers during this period must be carefully qualified. Pierre Guichard, for example, asserts that old, established Berbers became quickly Arabized and during the eleventh century those who had not lost total consciousness of their origin wanted to, and considered the new arrivals barbarians.(47) It is extremely difficult to evaluate the extent of the acculturation of Berbers at any one time. Literate Berbers wrote in Arabic, and therefore all our knowledge of Berber culture from literary sources is filtered through an Arabic screen. Then, many Berbers falsified their genealogies, adopting Arab tribal names in order to dissemble their true ethnic identity. Successful Berbers who made good in Arab society as judges or government officials had no need to disguise their origins, but many noble families resorted to falsification. Thus, the Zîrîd prince Habûs b. Mâksan, faithful to a Sanhâja tradition, claimed Himyarite descent, and the Aftasids, of Miknâsa origin, did likewise and used the Arabic tribal name Tujîbî.(48)

Nevertheless, acculturation did not necessarily lead to cultural fusion or to the lessening of social distance. 'Abd Allâh, the last Zîrîd king of Granada, whom Henri Pérès claims to have been totally Arabized, wrote that the Arabs had joined forces to fight the Berbers "owing to the hatred of their race."(49) The king may have been identical in culture to his antagonists, but he still perceived palpable social distance.

4. Assimilation of Neo-Muslims

To understand how and why the masses of indigenous Hispano-Romans came to form the mass of the population is the most difficult problem, but without attempting to do so we cannot hope to understand the dynamics of social organization in al-Andalus. The received view is that the Arab and Berber invaders, few in number, were absorbed into the indigenous social structure which continued unchanged, perpetuating considerable cultural distance from eastern Islam and making possible a continuity of interest with unconverted Christians. Thus Ramón Menéndez Pidal states that "Al-Andalus, so quickly made independent from the East, had Hispanified its Islam; the scant Asiatic and African racial elements had been almost completely absorbed within the indigenous elements, so that the great majority of Spanish Muslims were simply Ibero- [186]

Romans or Goths, reshaped (reformados) by Islamic culture, and who could easily enough come to an agreement with their brothers to the north."(50) This statement, which may be regarded as typical of traditional historiography, contains two kinds of questionable assumptions. The first concerns the directionality of cultural change. Pierre Guichard has demonstrated quite conclusively that far from ingesting the Arabs and Berbers into structures of native social organization, the muwallads themselves were assimilated to a measurable degree into the agnatically based social structure of the conquerors. Indeed, he shows that as a general rule endogamous, agnatic groups tend to expand precisely by ingesting exogamous elements, particularly women, without altering the basic structure of the family and, hence, of tribal organization.(51)

The second assumption concerns the rate of assimilation or fusion, no matter how the problem of directionality might be conceived. In general, fusion is thought to have occurred rapidly. Most commentators follow Lévi-Provençal in asserting that Arabs and Neo-Muslims had effectively merged into an "Andalusi Arab" group by the end of the tenth century. The problem of rate is associated with another question, that of the modality of the movement of assimilation or fusion. Fusion assumes a blending of cultural elements and a concomitant elimination of social distance between the groups involved. An example would be the fusion of Anglo-Saxons and Normans in medieval England, whereby the conquering minority was largely absorbed culturally by the conquered majority at the same time as social distance was being reduced (admittedly at a slower rate than the movement of acculturation). The result was a daughter culture, substantially different (as measured by the language, for example) from the two parents, partaking of elements of both. We will argue here that there was no fusion of Arabs and muwallads in tenth-century al-Andalus in the sense of a merging of cultures. The processes which took place can be summarized as follows:

1.The explosive phase of conversion, spanning the middle two quarters of the tenth century, resulted in the conversion of approximately eighty percent of the original indigenous population to Islam.

2.What Lévi-Provençal and others have perceived as fusion is not what acculturation theorists understand by that term (i.e., a blending of elements of two cultures, producing a distinctive daughter culture). Rather, the Neo-Muslims, while retaining certain indigenous and regional customs, institutions, vocabulary, and so forth, along the lines of other [187 Islamic provinces, acculturated massively to Arab norms. If language is used as the indicator, Romance/Arabic bilingualism, common in the early centuries, wanes with the fortunes of Christianity.

3. What historians have observed and mistaken for fusion was the swamping of the Arab minority by the mass of converts in the course of the tenth century, creating a different kind of society, an Islamic one in which indigenous peoples predominated, replacing the imperial state ruled by an Arab minority.

4. In spite of the Arabization of the indigenous masses, social distance was by no means eradicated, and the assimilation of the Neo-Muslims and their full acceptance by Arabs lagged substantially behind the movement of acculturation.

The mechanisms of the conversion process, we have already noted, were largely self-generating, depending upon the increasing frequency of contacts between Muslims and non-Muslims, owing to the incremental growth of the former group. The great mass of Neo-Muslims were initially rural people, although during the course of the country's rise to economic dominance in the western Mediterranean great numbers were attracted to the towns, where they swelled the ranks of the urban proletarian and artisan classes. To a measurable extent, then, there must have been economic and prestige factors which enhanced the attractiveness of conversion. Pastor contends that the Muslim sharîk (sharecropper) was better off than his equivalent in Visigothic times and that this obvious socio-economic differential was an impetus to conversion.(52)

This would accord well with Bulliet's generalization that lower-class people tended to convert first, in order to benefit from tax and other economic advantages which would accrue to them in a Muslim polity. On the other hand, the advantages of a change in juridical status would not immediately have been apparent to a Christian peasant who came into contact with Muslims only infrequently. To a city-dweller, avoidance of the jizya and the avenues of upward mobility available to the convert must have appeared obvious.

From a variety of indicators, one can piece together a picture of the steady conversion of the indigenous masses. Through the clientage system, Neo-Muslims were to some degree integrated into the Arabic tribal system.(53) The taking of Arab tribal names, a normal expression of clientage, was part of the process of Arabization and, moreover, the fact that Mozarab families arriving in León in the ninth century bore "Beni" names [188] indicates that at this early date even Andalusi Christians had adopted agnatic forms of kinship.(54) By the twelfth century, there is evidence of pervasive participation by muwallads in Arab tribal structure, attested to, for example, by cross-cousin marriages in the family of the muwallad lord ibn Mardanîsh (Martínez).(55) Even if agnatic kinship may not have spread uniformly among Neo-Muslims, certainly other, related aspects of Arabic social structure did. The social role of women, whose supposed "occidentality" is the subject of a particularly incisive critique by Guichard, is a case in point. Basing their conclusions on the numerical preponderance of indigenous peoples, Julíán Ribera and other Arabists supposed that the social structure of the numerically inferior Arabs was weakened and "Hispanized" by the massive intermarriage of Arab men and Hispanic women. Guichard shows not only that those women who married into Arab clans became part of the tribal structure and played social roles about as constricted as those of their eastern counterparts, but also that slave girls and concubines, through their rigorous training in Arab culture and mores, far from being agents of Hispanization were, in fact, agents of Arabization.(56)

The Arab rulers (again, in Guichard's view) used the agnatic kinship structure to good advantage in order to weaken the native aristocracy. By retaining most of the high governmental posts for themselves, the Arabs forced those Christians who wished to retain a measure of authority over their coreligionaries into the ecclesiastical hierarchy -- a biological dead end. By absorbing the rest of the native elite, particularly its women, into the clan structure, the indigenous upper class was quickly assimilated.(57)

The pattern of Neo-Muslim dissidence in the ninth and tenth centuries is similar to that of the Berbers, and provides a de facto measure of ethnic differentiation. There were three prominent foci of rebellion during the fitna which began in the reign of the Emir Muhammad: Mérida, where 'Abd al-Rahmân b. Marwân ibn al-Jilliqî ("son of the Galician"), whose father had been governor of Mérida under 'Abd al-Rahmân II, rose repeatedly (once in concert with Kurayb b. Khaldûn); the Upper March, where the Banû Qasî ruled in practical independence from Córdoba; and the mountains of Málaga, where the most important revolt erupted in 879 under the leadership of the charismatic muwallad leader, 'Umar ibn Hafsûn. In 899 he converted to Christianity, attracting Mozarab support but losing, in exchange, the following of most of his Neo-Muslim supporters, who had no intention of turning Christian. (Apparently his reasons for [189] conversion were opportunistic. He hoped to gain military support from Alfonso III of León.)

A number of aspects of ibn Hafsûn's revolt bear commentary. First, given the equalitarian promise of Islam, the muwallads, like the Berber khârijites, were fighting, in part, for their due share of power, as well as for the social and economic rewards guaranteed them by law. It is an error to view such revolts as "nationalist uprisings" or attempts at "Hispanic" independence; rather, they took place purely within an Islamic framework, as part of the growing pains of a nascent Islamic state. Seecond, ibn Hafsûn's cause was a social struggle fought by tenant farmers against Arab landlords. Third, the common cause made with Mozarabs was based on class interest -- the Mozarabs were also tenant farmers -- and no doubt the two groups shared elements of culture. But the chance for a lasting coalition of Neo-Muslims and Mozarabs was precluded by the fact that each group was embedded in Andalusi society in a different way, and the stakes in rebellion were quite different for each. Simply stated, Neo-Muslims had real access to power and could pact with other dissident Muslim groups, both Arab and Berber, toward the end of achieving it. Mozarabs could aspire to win relief from economic burdens deemed too heavy, but access to power was foreclosed to them. The cultural similarity and common origin of the two groups does not mean that their social aspirations or latitude for mobility were comparable.(58)

Muwallads were subjected to discrimination and racial stereotyping and reacted in characteristically defensive ways. In the early centuries Arabs called them "sons of slaves" or "sons of white women" and the social distance implied by such epithets was still palpable in the twelfth century. This is apparent in several observations of a defensive nature made by ibn Rushd, whom I assume to have been a muwallad. In a very interesting and revealing passage from his commentary on the Meteorologica of Aristotle, ibn Rushd invokes the ancient theory of climatic determinism to explain somatic differentiation between Andalusis and Arabian Arabs. The hair of Andalusis was straighter and less curly and their skin lighter than was the case with natives of Arabia. This is explained by the fact that, owing to the temperateness of the Iberian climate, the progeny of Arabs and Berbers residing there grew to resemble the natives physically (and, for similar reasons, the invaders began to study the sciences: since al-Andalus was similar in climate to Greece, the intellectual capacity of [190] the residents of both places was equivalent). Andalusis, therefore, who claimed Arab descent but who did not resemble Arabs somatically, had to have a rationale for that differentiation. Elsewhere, ibn Rushd stressed that nobility could be attained through achievement, as well as through birth, a position which is standard among groups seeking social recognition and equality.(59)

Difficulties of adjustment were also mirrored in expressions of ethnic solidarity among Neo-Muslims (and, on occasion, between Neo-Muslims and Christians). A muwallad judge from Huesca of the early tenth century, Muhammad ibn Sulaymân al-Ma'afirî (his grandfather had been a Ma'afirî client and so his tribal affiliation was fictive) was described as a "champion of muwallad ethnic solidarity" (shadid al-'asabîya li'l-muwalladin).(60) The notion of 'asabiyya (frequently translated "group feeling") is one of agnatic kinship and in this case is indicative of at least the style of agnaticism, if not the substance.

The shu'ûibiyya movement of the eleventh century was an effort by muwallad intellectuals to counter ethnic slurs which they continued to receive. In refutations of ibn García's Risála, Neo-Muslims were chided for their forging of genealogies to conceal their origin; mocked for their language (as "stammerers who have a speech impediment") and their blond coloration; decried as cowards in battle; derided for their disloyalty towards the Arabs and the lack of respect shown towards those who raised "your condition after it had been low"; and attacked as having Christian leanings, obvious since Christians were their "tribesmen." Ibn García's allegations on behalf of his ethnic group covered the same categories, only with positive valuation.(61)

To summarize the trend of relations among the three Muslim ethnic groups: through the ninth century armed dissidence was part of the process of adjustment through which competing groups tested each other's power, and by the manipulation and control of such conflict the state could react in such a way as to stabilize the polity. The late tenth century witnessed a drastic change in the ethnic balance of power, due, first, to the massive conversion of the indigenous population and, second, to an increment in the number of Berbers, both of which made untenable the survival of a unitary state dominated by Arabs. In the eleventh and twelfth centuries, distinctions between Arab and Neo-Muslim became increasingly blurred, although not forgotten, so that by the time of the Almoravid and [191] Almohad invasions, a more homogeneous culture, with regional idiosyncrasies parallel to those of contemporary Islamic states, had emerged.

5. Ethnicity in Christian Spain

The emergence of five distinct cultural zones in Christian Spain (Galicia-Portugal, León, Castile, Aragón, and Catalonia) did not in the high middle ages generate enough cultural distinctiveness to cause conflicts of an ethnic, as distinguished from a political, nature. During this period, people from all kingdoms spoke dialects which were mutually intelligible, or nearly so, and peoples from all regions seem to have intersettled in newly opened territory without conflict. Three Christian groups characterized by substantial ethnic differentiation -- primarily linguistic -- did evoke special responses, but were handled more in accord with the special statutes governing religious minorities as a more or less transitional measure to enhance peaceful intergroup relations during the period of acculturation. None of the three groups -- Basques, Franks, and the Mozarabs of Toledo -- suffered the long-term stigmatization that the religious minorities did.

There were numerous Basque settlements in the north of Castile, particularly in the region of Oca and Briviesca. We must assume substantial bilingualism throughout the period, and even the persistence of nuclei of monolingual Basque-speakers as late as the thirteenth century when (ca. 1235) Ferdinand III granted them permission to plead in Basque before the royal merino (judge).(62)

As they came down from the mountains, the Basques appear to have, on the whole, learned Romance very quickly and, since their participation in the formation of Old Castile and the early efforts of conquest and settlement was so much in evidence, seem not to have generated any hostility on account of linguistic differentiation.

The term franco was applied generically to anyone arriving from the north of the Pyrenees. The Chronicle of Sahagún specifies Gascons, Bretons, Germans, Englishmen, Burgundians, Normans, as well as folk from Toulouse, Provence, and Lombardy. But in fact most "Franks" were from the southern areas of France adjacent to Spain. These emigrants were primarily merchants and artisans who came to settle new towns and quarters springing up along the road to Santiago beginning in the late eleventh century. In these early settlements and extending well into the twelfth [192] century, the newcomers were spatially segregated and communally autonomous, preserving their traditional juridical norms. Thus in Estella, a burgum created solely for Franks by Sancho Ramírez of Aragón, the Fuero forbade any Navarrese to live there (1090). In the Fueros of San Cernín de Pamplona and Sanguesa (1122), it was stipulated that no one was to dwell among the Franks, "neither Navarrese, nor cleric, nor soldier, nor knight." In Belorado (1116) and in Sahagún (1152), there was one judge for Franks and another for Castilians. The rate of Frankish acculturation was a function of the distance of their settlements from the Pyrenean frontier and from the pilgrimage route. Thus, in Estella, the town ordinances were written in Provençal until the fourteenth century, while those who settled in the Duero Valley assimilated much more quickly.

The special privileges (not only juridical but economic -- the right to hold markets, monopolies over sale of provisions to pilgrims, tax exemptions, and so forth) accorded to the Franks induced ample resentment on the part of the indigenous Christians, provoking armed conflict as late as the thirteenth century in Pamplona. J. M. Lacarra has advanced the interesting hypothesis that embellishments of the legend of Charlemagne as liberator of much of northern Spain and pilgrim to Compostela, and tales of Frankish heroes come to serve St. James to free Spain from the Muslims, were promulgated by French jongleurs living in Frank communities along the pilgrimage road, in reaction to the hostility and envy that their presence generated among the natives.(63)

The most highly differentiated Christian minority was the substantial Mozarab population of the city and countryside of Toledo. By the time of the conquest of the town (1085) these Christians had become Arabic-speaking and they continued to speak and write personal documents in that language for two centuries or more.(64) In addition, they were highly acculturated to Andalusi customs and techniques (especially agricultural) which they continued to practice. Adding to their cultural particularity was the persistence of the Visigothic liturgy which the rest of the Castilians had abandoned in favor of the Roman rite. These were people who felt at home in an Islamic milieu, having preferred to remain in Toledo in early times when other Mozarabs had emigrated to Castile and León. From the very first moments of Castilian rule conflict erupted between the Mozarabs and the newcomers over the numerous estates abandoned by the Muslims. In the twelfth and thirteenth centuries there unrolled a [193] steady process of the impoverishment of Mozarab cultivators, as more and more land came under control of magnates and ecclesiastic corporations. The latter, under the influence of the intolerant Cluniac bishop Bernard and Rodrigo Ximénez de Rada, the primate archbishop, who was himself the principal buyer of Mozarab property in the early thirteenth century, fomented a segregationist policy under the cloak of religious nationalism.(65)

Ximénez de Rada's bias is symbolized in his coining of the semi-erudite etymology of the word Mozarab from mixti arabi,"(66) connoting the contamination of this group by overexposure to infidel customs, if not by migration.

On balance, then, the attitudes of Christian society toward ethnic minorities varied according to the degree of cultural distance, with religion carrying the heaviest charge, followed in importance by linguistic differentiation. In the case of the latter, social distinctions reinforced cultural boundaries and intensified ethnically based conflict. Thus Basques who settled among Romance speakers acculturated rapidly because they were not socially differentiated. For privileged foreign burghers and underprivileged Mozarabs, cultural boundaries remained high until social leveling could take effect, a process much protracted due to the mutually reinforcing influences of cultural and social factors.

Notes for Chapter 5

1. H. J. Chaytor, From Script to Print (New York: October House, 1967), p. 29.

2. On enaciados, see Castro, The Spaniards, p. 228.

3. Goitein, Mediterranean Society, II: 289.

4. Ibid., II: 311.

5. Eliyahu Ashtor, The Jews of Moslem Spain, vol. I (Philadelphia: Jewish Publication Society, 1973), p. 139; Goitein, Mediterranean Society, I: 251.

6. See Edward P. Colbert, The Martyrs of Córdoba (850-859): A Study of theSources (Washington, D.C.: Catholic University of America Press, 1962), pp. 31-33.

7. Yitzhak Baer, A History of the Jews in Christian Spain, 2 vols. (Philadelphia: Jewish Publication Society of America, 1961-66), I: 51; Rodríguez Fernández, La judería de la ciudad de León, p. 76.

8. Norma Mobarec Asfura, "Condición juridica de los moros en la alta edad media," Revista Chilena de Historia del Derecho, 2 (1961), 43. Burns, Islam under the Crusaders, p. 264 n. 39, notes a tendency in twelfth-century Castile to favor Christians in mixed cases.

9. Burns, Islam under the Crusaders, pp. 237-238.

10. Fuero de Calatayud: Muñoz, Colección de fueros, p. 462: "Et Christiano juret ad judeo, et ad mauro super cruce. Et judeo juret ad christiano in carta sua atora tenendo. Et mauro qui voluerit jurare ad christiano et dicat: Alamin catzamo et talat, teleta." In León, Jews had to take the oath by actually going to the synagogue and placing their hands on the Torah (Rodríguez Fernández, La judería de la ciudad de León, p. 75 n. 19); Fuero de Jaca, Molho, ed., pp. 56, 62.

11. Burns, Islam under the Crusaders, pp. 227-228, 265.

12. Goitein, Mediterranean Society, II: 277; Siete Partidas, 7.24.9; 7,25,10.

13. Siete Partidas, 7.24.7; 7.25.4.

14. Goitein, Mediterranean Society, I: 136-137; Siete Partidas, 7.24.10; 7.24.2; 7.25.1; Fuero de Jaca, p. 239 (no. 128); the Fuero Real of 1255 forbade Jews from reading anything contrary to their own religion, a good example of the dominant caste taking pains to strengthen the group boundaries of a subordinate caste (cited in Castro, The Spaniards, p. 546).

15. Goitein, Mediterranean Society, I: 142.

16. Ashtor, Jews of Moslem Spain, I: 58, 88, 155-227.

17. Andrew Handler, The Zirids of Granada (Coral Gables: University of Miami Press, 1974), pp. 72-77, 152-157. Cf. Lawrence Rosen's comments on the typical economic interdependence of Jews and Berbers in modern Morocco, "The Social and Conceptual Framework of Arab-Berber Relations in Central Morocco," in Gellner and Micaud, eds., Arabs and Berbers, p. 161.

18. Baer, Jews in Christian Spain, I: 50-51, 54, 56; Rodríguez Fernández, La judería de la ciudad de León, p. 66.

19. Goitein, Mediterranean Society, I: 199; Rodríguez Fernández, La judería de la ciudad de León, pp. 74, 82-84.

20. Ubieto, Ciclos económicos, pp. 7 (general hypothesis), 35, 42, 46, 78, 85, etc.

21. Baer, Jews in Christian Spain, I: 51; Mobarec, "Condición juridica de los moros," p. 46.

22. Burns, Islam under the Crusaders, p. 187.

23. Goitein, Jews and Arabs, pp. 125, 131; Ashtor, Jews of Moslem Spain, pp. 365, 388.

24. Goitein, Letters of Medieval Jewish Traders, pp. 9-10.

25. Ibid., p. 55.

26. Baer, Jews in Christian Spain, I: 63-64.

27. Ibid., I: 54.

28. Colbert, Martyrs of Córdoba, p. 301.

29. Vicente Cantarino, "The Ninth Century Cordoban Mozarabs: Did they Really Know Arabic?" typescript. Cited by permission.

30. Rodríguez Fernández, Monasterio de Ardón, passim. Gómez-Moreno, Iglesiasmozárabes, p. 15, also remarks that Mozarab women in León only rarely had Arabic names.

31. The pattern is common. See, e.g., Montagne, The Berbers, p. 7, on Berber bilingualism. Also, among the Sephardic Jews of northern Morocco of recent times it was typical to find bilingual or trilingual men married to women who spoke only Spanish.

32. Burns, Islam under the Crusaders, p. 389. Cf. Juan Torres Fontes, Los Mudéjares murcianos en el siglo XIII (Murcia: Academia Alfonso X el Sabio, 1964), pp. 10-11; in Murcia, the tribal majlis was also converted into a consejo de ancianos.

33. Dozy, Recherches, I: 37.

34. García de Cortázar, Epoca medieval, p. 111; also, p. 124, where he characterizes the Umayyad, Almoravid, and Almohad states all as fictitious entities, or superstructure. The "house of cards" characterization is that of Lévi-Provençal.

35. Ibn Khaldûn, The Muqaddimah, I: 332.

36. Isidro de las Cagigas, Andalucía musulmana: Aportaciones a la delimitaciónde la frontera del Andalus (Madrid: Instituto de Estudios Africanos, 1950), p. 55; Lévi-Provençal, España musulmana. Instituciones y vida social, p. 96.

37. See Lévi-Provençal, España musulmana hasta la caida del califato de Córdoba, trans. E. García Gómez, 2nd ed. (Madrid: Espasa-Calpe, 1957), pp. 74-75, 104, 132, 139-141.

38. Ibid., pp. 230-233.

39. Hady Roger ldris, "Les Birzâlides de Carmona," Al-Andalus, 30 (1965), 56-67; idem, "Les Aftasides de Badajoz," Al-Andalus, 30 (1965), 290.

40. See Idris, "Les Zîrîdes d'Espagne," Al-Andalus, 29 (1964), 52, 128, 130, 135; idem, "Les Aftasides," p. 290, and "Les Birzâlides," p. 59.

41. Emilio García Gómez, "Al-Hakam II y los beréberes según un texto inédito de ibn Hayyân," Al-Andalus, 13 (1948), 219-221.

42. E. Lévi-Provençal, L'Espagne musulmane au Xème siècle (Paris: Larose, 1932), p. 27; Idris, "Les Zîrîdes," p. 47; A. R. Nykl, Hispano-Arabic Poetry and its Relations with the Old Provençal Troubadours (Baltimore: J. H. Furst, 1946), p. 181; Rachel Arié, "Le costume des musulmans de Castille au XIIIe siècle d'après les miniatures du Libro del ajedrez," Mélanges de la Casa de Velazquez, 2 (1966), 64. Note, however, that Guichard (Al-Andalus, p. 305) rejects this line of argument, asserting that in the epoch of the conquest the turban was worn by Arabs, especially shaykhs.

43. Bosch Vilá, "Elemento humano norteafricano," p. 33.

44. Idris, "Les Aftasides," p. 54, and "Les Zîrîdes," p. 76

45. Idris, "Les Zîrîdes," p. 70 n. 20.

46. James T. Monroe, Hispano-Arabic Poetry: A Student Anthology (Berkeley: University of California Press, 1974), p. 160. A similar sentiment was expressed by the Umayyad prince al-Musta'in, who complained of the enslavement and Berberization of his dynasty (Nykl, Hispano-Arabic Poetry, pp. 72-73).

47. Guichard, "Peuplement de la région de Valencia," p. 123.

48. Bosch Vilá, "Elemento humano norteafricano," pp. 29-30; ldris, "Les Zîrîdes," p. 58, and "Les Aftasides," p. 278.

49. Pérès, Poesie andalouse, p. 259; Idris, "Les Zîrîdes," p. 43.

50. Menéndez Pidal, España del Cid, I: 77 (my translation). See comments by Glick and Pi-Sunyer, "Acculturation as an Explanatory Concept," pp. 144-145; and Guichard, Al-Andalus, p. 18.

51. Guichard, "Les Arabes ont bien envahi l'Espagne," p. 1,509; idem, Al-Andalus, pp.172, 192, 194.

52. Pastor, Del Islam al Cristianismo, p. 75.

53. Armand Abel's contention that clientage was not much in evidence in al-Andalus is unfounded; "Spain: Internal Division," in G. E. von Grunebaum, ed., Unity and Variety in Muslim Civilization (Chicago: University of Chicago Press, 1955), p. 214. See comment by Guichard, Al-Andalus, p. 330 n. 172.

54. Guichard, Al-Andalus, pp. 490, 442. The question of the kinship structure of the Hispano-Romans is unresolved. As Pierre van den Berghe has pointed out to me, the Hispano-Romans may well have been just beginning, at the time of the conquest, to move from a patrilineal to a bilateral family structure. Exposure to Arabs or conversion to Islam may well have arrested this movement or enhanced a shift back towards patrilineality (personal communication).

55. Guichard, Al-Andalus, p. 220.

56. Ibid., pp. 9, 148-149, 163, 172, 175, 178, etc.

57. Ibid., p. 197.

58. Ashtor, Jews of Moslem Spain, I: 108. On the dual structure of ethnic relations in Islamic Spain, see my article "The Ethnic Systems of Premodern Spain," Comparative Studies in Sociology, 1 (1978), 157-171.

59. On pejorative terms applied by Arabs to muwallads (itself a derogatory term), see Guichard, Al-Andalus, p. 213, and Abel, "Spain: Internal Division," pp. 214-215.

60. Lévi-Provençal, España musulmana: Instituciones y vida social, p. 102 n. 23. See also Guichard, Al-Andalus, p. 495, for other instances of muwallad 'asabiyya.

61. James T. Monroe, The Shu'ûbiyya in al-Andalus: The Risâla of Ibn García and Five Refutations (Berkeley: University of California Press, 1970), pp. 35, 36, 50, 51, 69, 76, 94.

62. Justo Pérez de Urbel, "Reconquista y repoblación de Castilla y León durante los siglos IX y X," in La reconquista española y la repoblación del pais, pp. 150-151.

63. J. M. Lacarra, "La repoblación del camino de Santiago," in ibid., pp. 225-228; idem,"La repoblación de las ciudades en el camino de Santiago," pp. 469-488.

64. See, e.g., the documents published by Angel González Palencia, Los mozárabesde Toledo en los siglos XII y XIII, 4 vols. (Madrid: 1926-1930).

65. Pastor, Conflictos sociales, pp. 199-268.

66. Colbert, Martyrs of Córdoba, p. 2I. The Chancellor Ayala later repeated the same theme: "mozárabes quiere decir christianos mezclados con alárabes." Cited by Ramón Menéndez Pidal, Orígenes del español, 2nd ed. (Madrid: Hernando, 1929), p. 434 n. 1.