The Making of a Mediterranean Emirate. Ramzi Rouighi
conditions.14 Clearly, intellectuals associated with the Ḥafṣids were not alone in producing notions of what Ifrīqiyā meant. However, by focusing on them, this book seeks to account for an important source of the inconsistencies behind both the historical and historiographic records.
Map 1. The medieval Maghrib.
State, Territory, and Region
Historians who have studied Ḥafṣid politics have assessed the power of the Ḥafṣid state by analyzing its control over its territory, the region of Ifrīqiyā. In this transposition of modern notions of state and territory, the natural expression of Ḥafṣid domination is assumed to be the homogeneous territory that fell between primarily political borders. For specialists such as Talbi, the Ḥafṣids ruled that territory from Tunis, their capital. This dynastic or state-centric perspective has the benefit of agreeing with the perspective of the sources. But it rests on one fundamental assumption, and that is the equivalence of the modern concept of “state” and the medieval Arabic term dawla.
The medieval word dawla is commonly translated it in English as “state” or “dynasty.” But dawla is attested in the earliest Arabic chronicles, and thus centuries before a notion of “the state” became established in European languages.15 Its meanings accrued and evolved in relation to widely different sociohistorical conditions and conceptions, and by the time Ḥafṣid chroniclers used the term, they were attributing to it meanings not all found in modern notions of the state.16
Certainly, Ḥafṣid authors applied the term dawla to the ruling family or dynasty. For example, in his Kitāb al-‘ibar (Book of Examples), Ibn Khaldūn (1332–1406) entitled the section on the Ḥafṣids “the dawla of the descendants of Abū Ḥafṣ” (dawlat Banī Abī Ḥafṣ).17 This usage suggests that the dawla was something the Ḥafṣids possessed or embodied collectively. In another example, the dawla was depicted as “Ḥafṣid,” as in “the Ḥafṣid dawla” (al-dawla al-ḥafṣiya). In this sense, “Ḥafṣid” was a quality the dawla possessed.18 This conception of dawla is, however, not the one most commonly found in the sources. Most frequently, dawla was applied to an individual ruler. This dawla began with the reception of the oath of allegiance and ended when the ruler died—or when he abdicated or was deposed. It is such a conception that the historian Ibn Qunfudh (d. 1407) applied when he wrote that the “dawla [of the Ḥafṣid al-Mustanṣir (r. 1249–77)] lasted twenty-nine and and a half years.”19 In this sense, dawla was a regnal period (mudda), not a kinship group or dynasty.20
While the “collective” dawla could continue to exist, in rare cases, without a Ḥafṣid emir, the more “personal” dawla could become extinct (inqaraḍat) while he was nominally in charge—although the end of the regnal period (dawla) had to await his actual death. This was the case of the dawla of Abū Ḥafṣ ‘Umar (r. 1284–95). “[His dawla] became extinct with the disappearance of its pillars (arkān).21 The first who died among its pillars was Abū Zayd ‘Īsa al-Fāzāzī—his family had means, leadership, and knowledge (‘ilm).”22 The conception of the dawla assumed here by the pro-Ḥafṣid historian Ibn Qunfudh differs from the aforementioned ones and refers neither to a dynasty nor to a regnal period. This dawla was imagined as a structure with weight-bearing walls or pillars. Significantly, this perspective envisaged a broader notion that included elite individuals outside the Ḥafṣid dynasty. Ibn Qunfudh also used the term dawla to describe an entity similar to the family. As the head of his dawla, the ruler was portrayed as keeping it in order, as a father would.23 This familial (dynastic) dawla was run like a business. When the ruler managed his dawla properly, peace extended beyond his narrow circle, benefiting the people (al-nās). Conversely, when the dawla was not in order (murattaba), everyone suffered. Interestingly, in spite of the multiplicity of meanings they lent a dawla, Ḥafṣid intellectuals never personified the concept. Their dawla did not raise taxes, build roads, or purchase goods. Moreover, it did not engage in relations with abstract entities such as tribes and cities. In this sense, it was not like the reified modern state.
Map 2. The tripartite division of the Maghrib.
Scholars have disagreed about whether the sources associate dynastic rule with a territory. Historian Mohamed Kably argued that the absence of personal names (nisab, sing. nisba) tied to a specific territory rather than a kinship group or a town suggests that the notion is simply an anachronistic imposition.24 In contrast, Dominique Valérian thought that the existence of terms pertaining to fiscal and administrative units “around” cities implied that such an association truly existed—in spite of the absence of an Arabic term that would have expressed that exact relation.25 Whether political authority and territory were associated can be debated, but the sources clearly do not make the connection through their use of the notion of dawla.
In his Histoire de l’Afrique du Nord, Charles-André Julien took the equivalence between dawla and state for granted. His map of North Africa in the thirteenth century illustrates the transformation of the rule of three dynasties into that of three states, each within a territory. Julien translated the tripartite political division of the Maghrib between the Ḥafṣids, ‘Abd al-Wādids (1236–1555), and Marīnids (1217–1465) into Ḥafṣid rule in Ifrīqiyā, ‘Abd al-Wādid rule in the central Maghrib, and Marīnid rule in the western Maghrib. As the title of his book states, this view establishes the modern nation-states of Tunisia, Algeria, and Morocco in the pre-modern past—the modern national territory of Tunisia being the successor of Ifrīqiyā.26 Unfounded and tenuous as it certainly is, this vision has enabled the wholesale nationalization of history.
The inattentive use of the modern notion of the state to interpret pre-modern texts has led historians such as Robert Brunschvig to conceive political history in terms of the changing power of the Ḥafṣid state. When he discussed the period that saw a number of Ḥafṣid emirs rule independently from the ruler of Tunis, Brunschvig naturally thought of it as a period of “fragmentation” and “disintegration.”27 He was not troubled by the fact that his point of view matched that of ideologues who favored a particularly Tunis-centric configuration of dynastic domination. Rather than account for them, Brunschvig dismissed the ideas of those who preferred autonomy as illegitimate, and presented the perspective of those who supported a strong ruler of Tunis as neutral and natural. Effectively, his state-centric approach depoliticized politics by cherrypicking types of change, and limited his analysis to the fluctuation between strength and weakness in the same Ḥafṣid state.28 This is a fundamental point that brings state-centric colonial scholarship and nation-state-based nationalist historiography in line with the staunchly pro-Tunis Ḥafṣid authors writing at the end of the fourteenth and in the fifteenth century.
In part, these difficulties arise because historians do not always, or fully, take stock of the changing composition of the ruling group, or do so but then take sides with one of the parties. That is why it is preferable to think of an oscillation of Ḥafṣid domination from a “regional” mode centered on Tunis to a “local” configuration with a multiplicity of autonomous Ḥafṣid capitals such as Bijāya. A change in perspective away from strength and weakness of a state is in keeping with this book’s lack of predilection for either of the two modes of domination. This approach also identifies the victory of the regional mode at the end of the fourteenth century as a political phenomenon that requires elucidation.
Characterizing a specific political configuration as regional explains why this book does not develop a theory of regionalization and then test whether it applied to Ifrīqiyā by mustering evidence. Its primary goal is not to use empirical evidence to demonstrate the nonexistence of a region—however one may define it;29