Version classiqueVersion mobile

Entre civitas y madīna

 | 
Sabine Panzram
, 
Laurent Callegarin

II.1. — Perspectivas generales

Early Medieval Urbanism in Ifrīqiya and the Emergence of the Islamic city

Corisande Fenwick

Texte intégral

  • 1 See his companion article, Kennedy, 1985.

1Kennedy’s (1985) «From Polis to Madina» article marked the start of an important chapter in the history of late antique and early Islamic scholarship. Overturning traditional assumptions about the Arab conquests and the end of the classical city, he argued that the decline of classical urban forms (particularly the disappearance of colonnaded boulevards, the privatization of the agora and the commercialisation of city-centres) were already visible as early as the 6th century and reflect the replacement of local civic authorities by centralized government. Equally importantly, he illustrated that the decline of classical urban forms did not mean the decline of the city: many cities in Syria-Palestine thrived and prospered under Muslim rule1.

  • 2 Thébert, Biget, 1990.

2A few years later, Thébert and Biget argued in an important article that Kennedy’s model held true for North Africa: the Vandal-Byzantine period was the point when the classical city «became medieval» and many cities prospered in Ifrīqiya (i.e. eastern Algeria, Tunisia and coastal Tripolitania) after the Arab conquest2. Since then archaeologists (and some historians) have devoted a considerable amount of attention to the question of the transformation of the classical city in late antique Africa and confirmed their arguments, but far less attention has been devoted to the question of how the city transformed after the Arab conquest in medieval Ifrīqiya. The aims of this paper are twofold: to briefly draw attention to a disjuncture between the way in which old and new cities are studied in Ifrīqiya by classical and medieval archaeologists, and to sketch out some key characteristics of urbanism in the post-conquest period based on the latest archaeological research.

The Polis to Madīna debate thirty years on: the medieval perspective

  • 3 Walmsley 2007; Avni, 2011; Id., 2014.
  • 4 Key excavations have taken place at Ramla, Fustat and Ayla.

3Let us begin by briefly considering the reception of Kennedy’s «From Polis to Madina» article in the Middle East for the Islamic period. Many of the details have been refined by several decades of high-quality fieldwork in Jordan, Syria and Palestine, but perhaps the most important challenge has been to Kennedy’s argument for a gradual drawn-out evolution from classical polis to Islamic madīna. The work of scholars like Avni and Walmsley has highlighted the huge degree of variability in the urban trajectories of old towns after the Arab conquests: some cities enjoyed massive growth and prosperity; other cities contracted but recovered and expanded in later periods; while a few cities especially in marginal steppe areas collapsed and were abandoned3. Rather than haphazard Muslim interventions into the urban fabric, the latest work at sites like Jarash, Tiberias and Scythopolis reveals planned residential and commercial quarters and the sympathetic and careful integration of new mosques, churches, markets and other buildings into the existing urban aesthetic under the Umayyad and early Abbasid caliphate. Excavations and analyses of Muslim urban foundations have demonstrated that the planned sites of Anjar, Qasr al-Hayr East and Samarra excavated in the 1980s were not rare exceptions: the seventh to ninth centuries saw the foundation of many new towns and cities laid out with regular plans and monumental architecture4. The latest work in the central lands of the caliphate, then, stresses variability in urban trajectory, the importance of new Muslim foundations and urban ideals and above all, the importance of looking at both new and old cities alongside one another.

  • 5 See Leone, 2007; Fenwick, 2013; von Rummel, 2016; Fenwick, in press a; Id., in press b.
  • 6 See Baratte this volume.

4In contrast, current scholarship on towns in the early Islamic period in North Africa has tended to fall into two distinct categories divided along disciplinary lines. The first, led by classical archaeologists, focuses on the «the Fate of the Classical City»: how and when were the monuments and public spaces of the classical Roman city transformed into their medieval form?5 Did classical towns flourish after the Arab conquest, or did they decline in size, population and prosperity? Scholars are in broad agreement on the first question: that the fifth-sixth centuries mark the point when the classical city «became medieval» in North Africa6. The latter question is more controversial. Debate centres around on how one interprets the evidence from several well-excavated towns that flourished in the Roman period and were abandoned at some point in the middle ages.

  • 7 Cressier, García-Arenal, 1998.
  • 8 On late antique Morocco, Villaverde Vega, 2001, on urbanization in Morocco, see especially Cressier(...)

5The second focus of scholarship, led by medieval archaeologists, might be described as the «Origins of the Islamic city» debate: what were the significant monuments and public spaces of the early Islamic city? What were the dynamics of urbanization under the caliphate and successor-states?7 Most archaeological research in recent years (in particular, the important work of Patrice Cressier) has focused on these questions in Morocco where Islamic archaeology is far more developed than elsewhere in North Africa8. Morocco, however, cannot be easily compared to Ifrīqiya: it was not under Byzantine rule (except perhaps the ports of Ceuta and Tangiers); it was barely urbanized on the eve of the Muslim conquests and in the 8th and 9th centuries, many towns and cities were founded under the new Muslim states that emerged. Morocco’s very different history therefore raises very different research questions about the medieval city than those raised in the densely urbanized lands of Ifrīqiya that had been under Byzantine rule.

  • 9 Lézine, 1971; Djelloul, 1995; Cressier et alii, 2001; Mahfoudh, 2003.
  • 10 Chabbi, 1967-1968.
  • 11 Louhichi, 1997.
  • 12 E.g. Cressier, Rammah, 2006. The eagerly awaited report of Cressier and Rammah’s excavations at Ṣab (...)
  • 13 See for example, Cirelli, 2001 on Leptis Magna; Ferjaoui, Touihri, 2005 on Zama Regia and Althiburo (...)
  • 14 Cressier, 2012, p. 119.

6In Ifrīqiya proper, medievalists have largely focused their attention on the monuments and organization of the caliphal foundations of Kairouan and Tunis and the palatial towns and the new monumental architecture (mosques, ribats and palaces) of the Aghlabid and Fatimid periods9. Excavations have been extremely restricted and largely limited to three palatial towns of Raqqāda10, and the later Fatimid foundations of Mahdiyya11, and Ṣabra al-Manṣūriyya12. It is only in the last few years that medievalists have begun to examine the post-conquest history of pre-existing towns in Ifrīqiya in detail13. As Cressier has pointed out, it is astonishing that only three sites have been the main focus of attention for over a hundred years — and indeed, even more remarkable that we still know so little about even these sites14.

  • 15 Fenwick, 2013, pp. 11-12.
  • 16 Mattingly, Hitchner, 1995.
  • 17 See Cirelli, 2001; Leone, 2007; E. Rocca and F. Béjaoui this volume for the value of re-examining o (...)

7The amount and quality of the surviving evidence for the 8th and 9th century phases of medieval cities in Ifrīqiya also plays a significant role and explains why the region has not seen the same amount of research as the Levant, Spain or even Morocco15. This region was densely urbanized in the Roman period and the colonial pre-occupation with this Roman heritage has had unfortunate consequences16. Almost all of the major excavated urban sites in North Africa were stripped down to their early imperial levels in the colonial period, making it difficult, though not impossible, to reconstruct any sense of the later history of these sites17. Late antique and medieval layers were destroyed, often without being recorded, until relatively recently and it remains the case that most projects at these sites are led by classical or Iron Age specialists with corresponding research agendas. Elsewhere, the critical evidence lies below modern cities: the major centres of early Ifrīqiya like Kairouan, Tunis (Roman Thuna), Tripoli (Roman Oea), Béja (Roman Vaga), Sousse (Roman Hadrumetum), Sfax (Tacapae) and Gafsa (Capsa) have all been occupied continuously. Almost nothing is known about their history before the ribāṭ-s and mosques built from the 9th century onwards.

Urban dynamics at the regional level

  • 18 Fenwick, forthcoming; Ead., in press.

8I have described elsewhere in more detail the evidence for regional patterns in Ifrīqiya and here summarize the key points18. The immediate changes brought to North Africa after the Arab conquest were minimal on many levels. First of all, it is clear that the amount of destruction was slight. Aside from the capital of Carthage that was stormed and sacked in 697/8 and never recovered, the majority of Byzantine towns were not destroyed or abandoned in the immediate aftermath of the Arab conquests. Second, it is apparent that the caliphate was not particularly intrusive and did not establish new settlements to the same extent as in the central lands of the caliphate.

  • 19 On bishoprics, see Maier, 1973.
  • 20 See Pringle, 1981 on Byzantine fortifications; Djelloul, 1995 on medieval fortifications.
  • 21 Cambuzat, 1982; Fenwick, forthcoming; cf. Kennedy, 2002; Id. 2011 and in this volume.

9Only two new towns — Kairouan and Tunis —were founded under caliphal rule (i.e. before 800), and as a result, the urban network of medieval Ifrīqiya continued to be based on that of Byzantine Africa (map 1, p. 206). With the exceptions of Kairouan and Tunis, both founded on, or near to, pre-existing settlements, it consisted of towns that had been occupied throughout the Roman period, and often many centuries before. Longevity was not a guarantor of success in itself: those towns that thrived in the middle ages share some common characteristics: they tend to be large, administrative centres that had usually obtained the juridical status of municipium or colonia and in late antiquity became bishoprics19. They are situated in strategic places on inland and coastal routes and often gained town walls or intra-mural fortifications in the Byzantine period that were restored or rebuilt in the eighth and ninth centuries — particularly those on the coast or in the Zab (ancient Numidia)20. Archaeology does provide evidence of some unfortified sites which continued to thrive in the early medieval period (e.g. Sbeïtla), but it is clear that most substantial towns were fortified in some manner by the 9th century at the latest. Perhaps the most significant determinant of urban success in the middle ages was the presence of a Muslim garrison which seems to have acted as a motor for economic demand in the same way as Kennedy has recently argued for the Middle East21.

  • 22 See Vitelli, 1981, pp. 15-17, pp. 24-39; Gelichi, Milanese, 2002; Stevens, 2016.
  • 23 Chelbi, Paskoff, Trousset, 1995; the ongoing Tunisian-British excavations at Utica have not found a (...)
  • 24 Fentress this volume; Potter, 1995.

10Alongside the survival of most of the large, administrative centres, the number of smaller towns seems to drop significantly by the 9th century. Many of the small towns of the northern Tunisian Tell, Numidia and the Tunisian coast seem to be abandoned or reduced to very little in the post-conquest period. Archaeological evidence suggesting 8th century abandonment or fragmentation largely comes from sites in the North, such as Uchi Maius, Chemtou and the old capital of Carthage22. We cannot, however, categorically date the failure of all abandoned towns to the early medieval period. Some towns in Ifrīqiya had been depopulated and largely abandoned well before the insecurities of the 7th century. The huge port-city of Utica, the first city to be founded by the Phoenicians in North Africa, was abandoned by the 5th century in response to the shifting channels of the river Medjerda and the silting up of the harbour23. The coastal cities of Meninx on Jerba and Cherchel in Algeria seem to have completely faltered in the 6th century24. These towns that failed in the prosperous 5th and 6th centuries provide a cautionary tale of making chronological assumptions about urban abandonment without careful, stratigraphic excavation.

Map 1. — Map of medieval North Africa (6th-9th century).

Source: Corisande Fenwick, based on Cambuzat, 1982

11Nonetheless, by the 9th century, the urban network had changed significantly from the urban peak of the early Roman period and its densely urbanized landscape of small and large towns. Carthage, the capital city of Africa under Punic, Roman, Vandal and Byzantine rule had been destroyed and replaced by two new cities: Kairouan, the new capital in the south and Tunis, the new harbour-city and district capital in the north. During late antiquity and the early middle ages, for whatever reason, many smaller towns do seem to have been reduced to villages or abandoned in the changing political and economic circumstances of the 9th century. In contrast, the large, prosperous, administrative centres of Roman Africa continued to be important centres in Muslim Ifrīqiya and benefited from the presence of soldiers, administrators and traders. Most changes in urban patterns, however, were not related to the direct intervention of the Muslim state and their speed and chronology varied for individual towns.

The new cities: Kairouan and Tunis

  • 25 The founding of a miṣr (pl. amṣār) usually adjacent to an existing town is an Arabic principle that (...)
  • 26 There is some dispute about the date and location of the first camps founded by the Arab armies, bu (...)

12When the Arab armies entered North Africa, they were already city-builders. During the conquests of Iraq and Egypt, a model had already been established for building amṣār (sing. miṣr), garrison cities to house the troops, at Basra (635), Kufa (639) and Fustat (641)25. It was not until 670 that ‘Uqba b. Nāfi’ founded the miṣr of Kairouan in central Tunisia as a base for his forces during the conquest period26. In 697/8 the Arabs finally defeated the Byzantines and took Carthage the capital; a few years later, Tunis was founded as a new harbour city to house the Muslim navy and act as a district capital in the north. Both were deliberately planned as regional centres and strategically placed in a region that was already rich in cities. In them can be seen a nascent Islamic urbanism.

  • 27 M’Charek, 1999, convincingly identifies this existing settlement as Iubaltianae and dismisses its e (...)
  • 28 Akbar, 1989; Whitcomb, 2007; Kennedy, 2010.
  • 29 Walmsley, 2007, p. 105.

13Precise information is much less plentiful for Kairouan and Tunis than for new foundations elsewhere in the caliphate. Nonetheless, the later Arabic texts have allowed scholars to reconstruct the key points in their development. The Arabic sources are confusing and contradictory on the early history of Kairouan. One tradition relates that ‘Uqba b. Nāfi’ built his new town ex novo in a plain covered by thick vegetation and home to reptiles and savage animals, while others suggest that the camp was established on, or on the outskirts of, an existing town traditionally identified as Qamuniya, and more recently as the road-side town of Iubaltianae27. We still lack a hypothetical reconstruction of the early layout of the site, nonetheless, it is clear that Kairouan follows the model of founding amṣār established several decades earlier at Basra, Kufa and Fustat where the founder’s responsibility was to lay out plots of land (hiṭaṭ), assign them to different tribal groups, build a congregational mosque and a governor’s residence (dār al-imāra)28. These towns were equipped with all the urban furniture required: mosques (faith, law, administration), palace (administration, justice, treasury) and markets (production, trade) and seem to have been laid out on a planned design. As Walmsley has noted for Syria-Palestine, this intersection of religious, administrative and commercial buildings shares some similarities with the urban ideals of Christian late antiquity, but adapted to fit the requirements of a Muslim population29.

  • 30 Sakly, 2000, p. 68.
  • 31 Ms of Ibn Abi Zayd, Mahfoudh, 2003, p. 75.
  • 32 Talbi, 1976.
  • 33 Northedge, 1994; Whitcomb, 1994; Hillenbrand, 1999.

14Kairouan was cut by large arteries, the most important of which was the simāṭ, a long straight street which cut the town into two from north to south and runs along the west face of the Grand Mosque where the first sūq was held (fig. 1, p. 208)30. The governor’s residence probably overlooked the simāṭ. Mahfoudh has drawn attention to an interesting account of a 9th century jurist who was asked to give his opinion about those merchants who were incorporating into their shops the colonnades on either side of the street, and blocking the passage to those on foot31. This is highly suggestive: we may here be seeing a colonnaded market street similar to those in Syria-Palestine. Unlike in the East, however, the street was probably not paved, but earthen and regularly flattened to keep it level32. The location of the mosque and surrounding streets suggests that the plan may have been on an orthogonal layout, with main axial streets similar to Ayla or Anjar, perhaps with irregular quarters and plots given to the different tribes as at Fustat33.

Fig. 1. — Great Mosque of Kairouan and the Simaṭ.

Photography: Corisande Fenwick

15Further construction took place under Ḥassān b. al-Nu‘mān who restored and expanded the Great Mosque in c. 689, though a more substantial building programme seems to have taken place under Hišām (723-742) whose governor Bišr b. Ṣafwān is also credited with repairing the mosque and giving it a minaret, building and laying out sūq-s on a long street along the west face of the Grand Mosque, as well as building fifteen reservoirs for the provision of water (fig. 2, p. 209). Here, we see clearly the role of the state in city-building: on the caliph’s instructions, the city was provided with a working infrastructure, including markets, a water supply (cisterns, aqueducts), as well as a congregational mosque for the believers.

Fig. 2. — Cisterns of Sidi Dahmani, Kairouan.

Photography: Corisande Fenwick

  • 34 Mahfoudh, 2003, pp. 19-39. Apparently a ditch was cut in 770 during a siege of Kairouan that was so (...)
  • 35 Ibid., p. 28.
  • 36 Id., 2008.

16Immediately after the Abbasids reconquered North Africa in 761/2, the new governor built defensive walls in pisé34. At this point, this was certainly for pragmatic reasons, but it may also be as Mahfoudh has suggested, a symbol of Arab authority35. The fortification of the amṣār seems to be an Abbasid innovation. Al-Saffāḥ, for example, authorized the governor of Fustat to build in 752 a new fortified town «al-‘Askar», whilst al-Manṣūr gave ramparts and ditches to the Iraqi amṣār of Kufa and Basra in 771. In 771, the Muḥallabī governor Yazīd b. Ḥātim demolished all but the miḥrāb of Uqba and rebuilt it. At the same time, he reorganized the principle sūq-s and regulated the transactions taking place in them. By the end of the caliphal period, Kairouan possessed a Great Mosque alongside dozens if not hundreds of smaller public and private mosques, a dār al-imāra, several distinct cemeteries and sūq-s organised by the two main neighbourhoods of the Quraysh, the Meccan tribe of the Prophet, and the Anṣār, the families of the Medinan tradesmen who supported the Prophet36.

17Tunis was not a miṣr, but like Kairouan it was built as a Muslim city. According to the sources, on 9 March 699 Ḥassān b. al-Nu‘mān founded the new city and arsenal at the base of the lagoon, on the outskirts of Tunes, a minor town and bishopric in the 7th century. As well as building or dredging a canal between the sea and lagoon, he imported Coptic labourers from Egypt to build an arsenal and ships for a new naval fleet. Replacing the Byzantine capital, Carthage, which had been captured and destroyed in 697/8, it took on the role of capital of northern Tunisia and the biggest export port.

  • 37 Lézine, 1971, pp. 141-154. Bishops attended Councils in 411 and 533: Mesnage, 1913, pp. 164-165.
  • 38 Al-Bakrī, Description de l’Afrique septentrionale, 81-82.
  • 39 Ibnawqal, Configuration de la Terre, 70.
  • 40 Mahfoudh, 2003, pp.173-209.

18Little is known about the history of Roman Thunes, which is barely mentioned in the classical sources: it seems to have been a fortified town and a bishopric protecting Carthage from its south37. This is a rare example of a town seeming to be totally abandoned at the moment of the conquest. Al-Bakrī gives us the fullest account and tells us that the inhabitants of Thunes had fled when Ḥassān b. al-Nu‘mān captured it, and then established a garrison there and built a mosque38. Ibn Ḥawqal is more specific and tells us that «after the Muslims had built new constructions there, renovated the gardens and walls, it was called Tunis»39. Finds in the western part of the madīna, especially in the vicinity of the Zaytūna mosque, support the reading that that the medieval town was indeed founded in the abandoned Roman centre40. Of the city itself, little is known beyond the various construction dates and builders given for the Zaytūna mosque in Tunis ranging from Ḥassān b. al-Nu‘mān in 703 to 731, though an early date seems most likely.

  • 41 Marçais, 1954, pp. 9-22.
  • 42 See Johns, 1999, pp. 109-110.
  • 43 Leone, 2007.

19Kairouan and Tunis represent huge investments of labour, materials, money and time which no other North African city received under the caliphate. The construction of a large congregational mosque in the centre of both towns, in particular, would have articulated a very different visual identity from the urban landscapes of old cities. Early congregational mosques only seem to have been founded in Kairouan (c. 670), Tunis (early 7th century) and perhaps also at Tripoli during the first raids41. We know almost nothing about the appearance of these mosques before they were rebuilt under the Aghlabids in the 9th century though presumably they conformed to the architectural template of congregational mosques in the central lands of the caliphate, which had been established by 70042. We might imagine that the mosques of Tunisia were built upon similar lines as the near-contemporary Umayyad mosques constructed in the East under the patronage of al-Walīd I (705-15), such as the Aqṣà mosque in Jerusalem or the Mosque of Damascus. Filled with new architectural forms like the mosque, Kairouan and Tunis would also have been visually distinctive because of what they lacked of the old urban order. Occupied for centuries, other cities were a palimpsest of North Africa’s complicated history, protected by defensive walls and intra-mural forts and filled by churches as well as monumental complexes such as theatres, baths and temples, many of the latter already in ruin and repurposed for other functions43.

The old cities

  • 44 Pringle, 1981. On the shifting ideal of the city in the 6th century, see Saradi 2006.
  • 45 Leone, 2007, pp. 166-279. See also: Mahjoubi, 1979; Février, 1983; Thébert, 1983; Pentz, 1992; Lepe (...)

20On the eve of the Muslim conquests, North African towns looked very different from those of the early Roman period. Major structural changes in towns had already occurred during the sixth and seventh centuries when a growing emphasis on security resulted in the construction of town walls and intra-mural fortresses under the Byzantines44. The new monumental buildings of the 6th and 7th century were churches and fortifications. At the same time, formerly public spaces like the fora were often adapted for other uses, some industrial and agricultural activities moved into towns and peripheral zones of towns were often abandoned45.

  • 46 For bibliography, see Gelichi, Milanese, 2002; Fenwick, 2013; von Rummel, 2016.

21Urban developments in the inherited towns of Ifrīqiya after the Arab conquest are much more obscure, in part because archaeologists find it so very difficult to date early medieval activity. Even so, the number of known cases is beginning to increase rapidly through a combination of new archaeological work at key archaeological sites (e.g. Henchir Douamis [Uchi Maius], Henchir el Faouar [Belalis Maior], Sbeitla [Sufetula], Haïdra, Althiburos, Chemtou, Jama [Zama Regia]) and the re-evaluation of earlier excavation reports (e.g. Tébessa, Lebda, Carthage)46. This work reveals the great degree of urban variability in the early middle ages. As scholarship develops on the question of the survival and re-occupation of classical sites, it will therefore be important to distinguish between those towns that continued to function as such in the middle ages; those towns that may have contracted but recovered and expanded in later periods; and those towns that collapsed and were re-occupied but never regained their urban status.

Fig. 3. — Plan of Sbeïtla.

Modified from Duval, 1980, with additions from Bejaoui, 1990 and 1996

  • 47 For a full account of medieval Leptis, see Cirelli, 2001.
  • 48 Fenwick, 2013, pp. 21-23, for a re-analysis of Duval, 1982; Id. 1990, model of a city reduced by th (...)
  • 49 Duval, Baratte, 1973; Béjaoui, 1996, pp. 38-39.
  • 50 Duval, 1982, p. 625; Id., 1999.

22Here I wish to focus on those towns that seem to have both survived and continued to maintain some form of urban status. Many of the larger towns that thrived in the Middle Ages seem to have been the same size as they were in the Byzantine period. Cirelli has ably pulled together the evidence for medieval Lebda (Lepcis Magna) and shown that the town continued to be the same size as in the 7th century, comprising an enclosed settlement of 28 ha around the port area, perhaps with some pockets of occupation outside the walls in the market, chalcidicum, circus, and Hunting Baths47. Even with a silted-up harbour, Lebda remained a sizeable centre into the 11th century and continued to trade, producing and importing globular amphorae and olive oil on some scale. The same holds true for other towns like that of Sbeïtla (anc. Sufetula) in central Tunisia (fig. 3)48. Although we do not yet know the full spatial extent of early medieval Sbeïtla, it is clear that the residential area around the old forum, and the southwest area remained occupied into the 9th century. Fortified dwellings or «fortlets», occupied in the seventh-ninth centuries, line the southern entrance to the town49. All the churches (Basilicas I, II, IV and V) continued to be used after 650, and the latter two continued in use until the 10th/11th century50. By the 9th century Sbeïtla was still a thriving and ordered settlement on an orthogonal plan with churches, workshops, presses and a working drainage system.

  • 51 Mahjoubi, 1978.
  • 52 Fentress, 1987; Amara, Fentress, 1990; Fentress, 2000; Ead., 2013.
  • 53 Biskra: al-Bakrī, Description de l’Afrique septentrionale, 111; Baghai: Ibnawqal, Configuration d (...)

23At other towns, the presence of Arab-Muslim garrisons seems to have been a key factor. While they often may have simply have taken over the Byzantine forts that many towns already possessed, at other towns they made an imprint on the urban organization. At Henchir el-Faouar (anc. Belalis Maior) in Tunisia (fig. 4), for example, an 8th-century quarter with a small fort and courtyard houses laid out on an orthogonal plan was excavated in the 1960s51. The new quarter was constructed on the northern fringes of the town, in a peripheral area that probably served as a cemetery in the 7th century, as is suggested by the razed foundations of a cemetery church below the fort. Next to it are several houses neatly laid out, following a courtyard plan that was not previously known in North Africa. Fentress has convincingly suggested that this type of courtyard plan is Arab in origin, on the basis of its similarity with 6th-century Arab housing from Umm el-Jimal: it appears at several other sites outside the walls in the eighth and ninth centuries52. The new settlement at Henchir el-Faouar seems to have been military in nature, but new suburban extensions may also have served the growing non-military Muslim community particularly in the ninth and tenth centuries. At other sites, unknown archaeologically, geographers like al-Bakrī and Ibn Ḥawqal describe the new suburbs that grew around major garrison centres such as Biskra, Bagai and Tobna53.

Fig. 4. — Satellite photo of medieval settlement at Belalis Maior.

  • 54 See Laporte, 2015 for a more detailed overview of the late antique and medieval city, supplemented (...)
  • 55 See Foucher, 1964, p. 341, though the argument seems to be largely based on the presence of sculptu (...)
  • 56 Lézine, 1956.
  • 57 Talbi, 1966.
  • 58 Lézine, 1956, pp. 44-45, suggests that the wall follows Byzantine foundations, a suggestion that is (...)
  • 59 Carton, 1907; Djelloul, 1995.
  • 60 On the mosque see Creswell, 1989, pp. 353ff.; Marçais, 1954, pp. 23ff.
  • 61 Lézine, 1971, p. 104.
  • 62 Baratte et alii, 2014, pp. 221-222.

24Much less is known about the most important inherited towns of Ifrīqiya like Béja, Sousse, Sfax, Tobna and Gafsa which served as important military and administrative centres in the early medieval period. The coastal town of Sousse is a good example and worth elaborating on (fig. 5, p. 214). Little is known about the layout of Byzantine Justinianopolis (Roman Hadrumetum), the capital of the province of Byzacena, but it was certainly a prosperous fortified port-city when it was conquered by ‘Uqba b. Nāfi’54. It continued to be a military and regional capital under the Umayyads and Abbasids and in the late 8th century, a fort (ribāṭ) was built to guard the interior port, near or on, what was presumably the Roman forum. This is the oldest known ribāṭ in Ifrīqiya and may have been built on top of the foundations of an ancient church55. The ribat is now dated on archaeological grounds to the late eighth century and it is generally assumed that it was built by the Muhallabid governor (r. 772-788)56. Under the Aghlabids, Sousse became the main naval base rather than Tunis and was used as the departure point for the first expedition with 99 ships to Sicily in 82757. Around this point, we begin to see major investment in its infrastructure by the Aghlabid emirs: a manār (watch-tower) added to the ribāṭ; the small mosque of Bu Fatata in 838-41; a Kasbah built in 844-5; the grand mosque in 851, and finally in 859, a massive defensive enceinte was built or rebuilt (perhaps following the line of Byzantine walls) that encircled 32 ha and encompassed the entire city including the arsenal and interior port (fig. 6, p. 215)58. Outside the walled city, several small fortifications were built to protect the main route ways59. It is worth emphasizing that before the construction of the great mosque in 851, the inhabitants of Sousse had access to several mosques. The earliest is the prayer hall (36 m × 7 m) on the upper floor of the ribāṭ dated by inscriptions to 206/821, and presumably there was one in the earlier ribāṭ60. The second is the small nine-bay mosque of Bu Fatata with a façade of three horseshoe arches founded in 838-841 under Abū ‘Iqāl al-Aġlab (r. 223/838-226/841) and his freedman Hiḍr. This mosque in the centre of town presumably was open to all, but only consisted of a small prayer room (10 m × 10 m) and a narthex. A further small prayer room (4 m2) in the tower of Halaf (851) was surely restricted to the use of the garrison61. We do not have any further references to mosques in Sousse, though they may have existed. As for churches, these surely continued to exist (apart from the one that was destroyed by the construction of the ribāṭ), but the other identifications outside and inside the walled city are all disputed62.

  • 63 Février, 1965; Fentress at alii, 1991.

25The variability of urban trajectories in this period is striking. If all these examples suggest continuity, Sétif provides a clear example of urban contraction in the seventh/eighth centuries followed by expansion in the ninth-tenth centuries. Continuously occupied, the excavators suggest that the town retreated into the Byzantine citadel in the early medieval period63. Excavations in the 1980s found continued ephemeral medieval housing to its west, whilst the area north of the citadel seems to have been abandoned after the Roman baths were destroyed. In the 9th/10th century, this latter area seems to be re-used as a market space, surely a mark of its liminal location on the edge of the town, before being transformed into a new residential quarter that expanded gradually over the course of a century.

  • 64 Béjaoui, 1996.
  • 65 Jones, 1983; E. Rocca and F. Béjaoui this volume.

26The effect of the Muslim conquest on the ancient street plan and on urban utilities (especially the water system) has been much debated. In most North African towns, the basic urban street grid plan did not change significantly from Byzantine to early Islamic times. We can see this very clearly at sites like Sbeïtla where the gridded street system survived intact into the middle ages, with few signs of encroachment. New medieval houses followed the lines of the Roman insulae. The street drainage system also continued to function, and was repaired in the post-conquest period64. At other sites such as Tocra (in Cyrenaica) or Haïdra, although the wide paved streets may have been narrowed by the addition of shops, the main arterial Roman streets were in continuous use during early Islamic and later medieval times65.

Fig. 5. — Plan of Sousse (9th c.).

Draw: Corisande Fenwick

  • 66 Mahfoudh, 2003.
  • 67 Mahfoudh, Baccouch, Yazidi, 2004, p. 11; Mahfoudh, 2008.

27Water supplies continued to be maintained and improved. As described above, the governors of Kairouan invested a huge amount of money into providing the people of the town with sufficient water for ablutions, cooking and for their animals by cutting wells and building large storage basins66. The cisterns of Sidi Dahmani built under Hišām (in the same location where wells had been cut in the 660s) seem to have been fed by a diversion from the oued Merguellil and a ceramic channel has been found nearby (fig. 2, p. 209). When Hišām renovated the great mosque, a vast cistern was also cut in the courtyard which still exists today. The small oratories in each quarters were also given cisterns (and usually ḥammām-s) even in this period. The slave-trader turned missionary, Ismā‘īl b. ‘Ubayd who arrived in the early 8th century gave his oratory (the Zaytūna Mosque) a cistern for the use of the town’s inhabitants67.

Fig. 6. — Sousse showing great mosque, port and walls taken from the ribāṯ.

Photography: Corisande Fenwick

  • 68 Béjaoui, 1996.
  • 69 Mahfoudh, 2003.

28We know far less about the situation at other towns, where presumably pre-existing rain-fed cisterns continued to be maintained and used, even when aqueducts were no longer in use. At Carthage, on the south flank of the Byrsa hill, in addition to a large number of cisterns, there is piping leading down the hill, though little is known about it, and in the Bordj Djedid region, an aqueduct was found which may have been dredged out in the medieval period. At Sbeïtla, too, a pipeline was inserted along the cardo in the medieval period68. At major cities, these works were often carried out by the governor or emir. Thus in the 9th century, an account of al-Mālikī indicates that in the 850s, the Sofra cistern, a vast reservoir with a capacity of 3,000 m3 dating to Byzantine period was being used by the Aghlabid emir69.

  • 70 Jones, 1984; Thébert, 2003.
  • 71 Bel, 1913; Poinssot, 1958; Fentress, Limane, 2010.
  • 72 Al-Bakrī, Description de l’Afrique septentrionale, 26.

29A working water supply and drainage system was certainly needed for bathhouses. The demise of the large bathhouse was an earlier phenomenon, but at Tocra, at least, the main baths existed in a reduced version at least until the 8th or 9th century, as attested by the Kufic inscription on the threshold70. New baths dating to the eighth or ninth centuries have been uncovered at Dougga (fig. 7, p. 216), Pomaria-Agadir (next to the mosque) and Volubilis (near the wādī)71. Our poor understanding of baths and bathing patterns in the early medieval period is particularly problematic given the importance of ritual ablutions in Islam and the high numbers of baths (ḥammām-s) noted by the later Arab geographers in individual cities. Al-Bakri, for examples, notes that by the mid-11th century at Kairouan, there were no fewer than 48 ḥammām-s72.

  • 73 Pringle, 1981.
  • 74 Cambuzat, 1982.
  • 75 Goodchild, 1967; Jones, 1983.
  • 76 Fenwick, 2013, pp. 26-27.

30Most of the inherited cities were heavily fortified with defensive town walls and/or a fort or citadel in the centre of the town. These fortifications are almost exclusively Byzantine-built, mostly in the 540s in the immediate aftermath of the Byzantine reconquest73. Town walls did not surround the entirety of the town, only the military and economic core. The fortresses were often taken over to house Arab garrisons74. Firm archaeological evidence for the Arab takeover of forts comes from the large intra-mural fortress at Tocra hastily built by the Byzantines in response to the first Arab raids in the 640s75 (fig. 8, p. 217). A second phase consisting of new buildings and a bath complex of Eastern rather than North African plan similar to that of the 8th-century bath at Volubilis surely reflects its takeover by a Muslim garrison. Less securely dated instances of this phenomenon are recognizable in later additions of mosques, repairs and medieval occupation to Byzantine fortresses, as at Ksar Lemsa, Béja, Dougga, Djaloula, Zana, Tigsis, Ksar Belezma, Tobna and Bagaï, for example76.

Fig. 7. — Medieval baths at Dougga abutting the exterior of the Byzantine fort.

Photography: Corisande Fenwick

  • 77 Mahjoubi, 1978.
  • 78 Al-Bakrī, Description de l’Afrique septentrionale, 56-61.

31The Umayyads and Abbasids also built new fortifications in the 8th century. Initially there seems to have been little need to build new fortifications and the only known early 8th century example is the small fort (27.20 × 38.80 m) found on the outskirts of Belalis Maior (Henchir el-Faouar), dated by a coin between 709-71777. The Abbasid recapture of Ifrīqiya in the 760s was followed by a reorganisation of the army and an investment in building monumental fortifications anew or strengthening the existing defenses at the biggest towns. According to al-Bakrī, Kairouan was given a wall in pisé by Ibn al-Aš‘āṯ, the first Abbasid governor78. Tobna which became capital of the Zab under the Abbassids was also given a wall on the instructions of the caliph al-Manṣūr, which were then reconstructed by one of the Muhallabid-s in 768.

  • 79 Djelloul, 1995; Hassen, 2001.
  • 80 Pringle, 1981.
  • 81 Carton, 1905.

32Towards the end of the 8th century, ribāṭ-s began to be built along the Tunisian coastline to protect coastal towns and villages from marauding Byzantine fleets in the Mediterranean79. These imposing new fortifications look very different to the forts built by the Byzantines80. The earliest of these seems to be the ribāṭ at Sousse discussed above which has a square plan (38 × 38 m) with four outer towers and an interior consisting of a central courtyard surrounded by rooms and an upper floor that housed a prayer hall. Around the same time, a ribāṭ on this distinctive plan was built at Monastir in 796 on the order of the governor of Ifrīqiya, Hartama b. A‘yan (fig. 9, p. 218). Coastal forts continued to be built at some of the old towns in the early 9th century under the Aghlabids. Thus Qaṣr Ziyād (826) was built on the site of Ruspae, whilst the ribāṭ at Ras Capioudia (Caput Vada) was built upon Byzantine foundations and guarded the harbour entrance81.

Fig. 8. — Plan of early medieval Tocra showing location of Byzantine fort.

Draw: Corisande Fenwick

  • 82 Talbi, 1966.
  • 83 According to Ibnaldun, Histoire des Berbères, t. 4, p. 429; also Talbi, 1966, p. 251, under Abū-I (...)
  • 84 Djelloul, 1995.

33A second phase of fortification dates to the middle of the 9th century, when the Byzantines responded to the intensification of Aghlabid raids on Sicily, Italy and Mediterranean islands by descending on the African coastline82. The Byzantines seem to have focused on sub-urban zones and unprotected coastal towns, notably in the Sahel and Cap Bon rather than great cities. Although it is very difficult to date ribāṭ-s in the absence of excavations, it seems that many ribāṭ-s were built along the coast in this period so forts were close enough to signal to one another in case of distress83. At least twenty ribāṭ-s can be firmly dated to this period, including the single storey ribāṭ at Lamta (860), the isle of Ghedamsi (871)84. Sousse and Sfax were also given walls in the period (though these may also have been restorations of Byzantine defences).

  • 85 Ibid., p. 38.

34By the mid-9th century, then, many of the coastal towns had gained new fortifications which often changed their appearance significantly. Whether or not they were surrounded by walls, towns were often defended by a series of small forts or isolated towers that governed their approaches (e.g. Kairouan, Tunis, Sousse, Sfax). The multiplication of these isolated works outside towns, as well as along the main transportation routes seems to be the most marked trait of Aghlabid fortifications85. One wonders whether some of the small watchtowers converted from arches or other earlier buildings often identified as Byzantine at inland towns like Mactar and Haïdra may in fact be medieval additions.

Fig. 9. — Ribāṯ of Monastir.

Photography: Corisande Fenwick

  • 86 Handley, 2004. On Christianity after the Muslim conquest, see Mahjoubi, 1966; Talbi, 1990; Savage, (...)
  • 87 Duval, 1982, p. 625.
  • 88 Talbi, 1990, p. 319.
  • 89 Church III in Gui, Duval, Caillet, 1992, p. 25.
  • 90 Gauckle, 1907, p. 794.

35The final element we can consider is religious. Churches continued to be a dominant element in the urban landscape, though it has proven difficult to identify this archaeologically86. Some churches certainly did retain their Christian function for centuries after the conquest, as at Sbeïtla, where Basilicas IV and V seem to have continued in use until the 10th/11th century, though others were given over to secular purposes at some point87. Christians were not simply continuing to use existing churches, but building new ones. Thus in the late 8th century Qustas (Constans) was granted permission to build a church in Kairouan88. The majority of churches, however, were repurposed at some point, such as the church at Belalis Maior, transformed into an oil pressing facility, or the church at Tipasa transformed into a market at some point in the medieval period89. Others still were dismantled for their materials, though 10th/11th-century judicial texts suggest that this was only supposed to happen to abandoned churches. Rarely in this period were churches transformed into mosques, as at El Kef and (reportedly) the Zaytūna mosque at Tunis90.

  • 91 Marçais, 1954, pp. 29-30.
  • 92 Dunbabin, 1978, pp. 9-22.

36The religious impact of the mosque on cityscapes varied greatly from city to city. Outside Kairouan and Tunis, there is little firm archaeological evidence for mosques in the Umayyad or Abbasid periods, though they would certainly have been built wherever Muslim troops were stationed. Mosques or prayer rooms have been identified inside the citadels of Belalis Maior, Bagaï, Tobna as well as the late 8th-century ribāṭ-s of Sousse and Monastir91. In the towns proper, there is little textual or archaeological evidence for mosque construction before the Aghlabid period outside the provincial capitals of Kairouan and Tunis92. Initially, then, it seems possible that mosques were largely confined to fortresses for the Muslim garrison troops, and to the provincial capitals, where we might expect large numbers of incoming Arabs to live.

  • 93 Mila Carthage: Whitehouse, 1983; Haïdra: Baratte, 1996, pp. 151-153.

37It is not until the start of the 9th century that archaeology and texts provide evidence of large-scale foundations of mosques across Ifrīqiya by the emirs and local pious individuals. As well as the congregational mosques of Sousse (851), Sfax (850) and Monastir (9th century) and the many small neighbourhood mosques dated by texts and inscriptions, we can add a few archaeological examples from Mila, Carthage and perhaps Haïdra, where a small medieval building with columns and a trapezoidal plan known as the «batiment à colonnes», has also been uncovered inside the citadel, and is tentatively interpreted as an early mosque, though no date is given93.

38The study of the towns of early medieval Ifrīqiya and their transformation from late antiquity is still in its infancy, and lags behind medieval studies in, al-Andalus. Perhaps the most important conclusion to draw is that the historical trajectories of towns varied enormously whether they were the old, inherited classical towns or new Muslim foundations. In many cases, the fate of cities were decided by political decisions: those cities that gained garrisons of Arab troops or served as administrative centres thrived and prospered in the middle ages. This was not simply because they became centres of Muslim activity, but because in the early Islamic state, the government and military were the most important generators of economic activity.

39Despite a marked lack of investment in many of the old towns in the eighth and early ninth centuries, it is clear that the Umayyad and Abbasid armies moving in did have a significant effect on urban organization. The new foundations of Kairouan (670) and Tunis (c. 705), both built on or near existing settlements, mark the most obvious break with the old order. Established as new Muslim cities to house the Muslim soldiers and ever-growing Muslim community, they included new types of buildings including mosques and the governor’s residence (dār al-imāra) that had never been seen before in North Africa. At other sites, the takeover of forts and the addition of extra-mural quarters to house the new Muslim community increased the size of other sites and introduced new forms of house plan.

40The late 8th and 9th century mark a more significant change for North African cities with new ribāṭ-s, town walls, congregational mosques and oratories being constructed on a large scale in many of the old towns as well as the new towns of Kairouan and Tunis. A new model of Islamic urbanism also appeared in the new palatine towns of al-‘Abbāsiyya (800) and Raqqāda (876) established by the Aghlabid emirs which I have not discussed in this paper. At the same time, the Roman and late antique heritage continued to strongly shape the urban form of many of the old towns of Ifrīqiya and differentiate them from the new Muslim foundations. Churches and Byzantine-built forts and town-walls were still dominant elements in many cityscapes, as were the Roman streets and water supply and storage system that underpinned their ability to function as urban centres.

Notes

1 See his companion article, Kennedy, 1985.

2 Thébert, Biget, 1990.

3 Walmsley 2007; Avni, 2011; Id., 2014.

4 Key excavations have taken place at Ramla, Fustat and Ayla.

5 See Leone, 2007; Fenwick, 2013; von Rummel, 2016; Fenwick, in press a; Id., in press b.

6 See Baratte this volume.

7 Cressier, García-Arenal, 1998.

8 On late antique Morocco, Villaverde Vega, 2001, on urbanization in Morocco, see especially Cressier, 1992; Id., 1998; Bonne, Benco, 1999. See Ettahiri, Fili, Van Staëvel, 2013, pp. 157-160 for a brief history of developments in Islamic archaeology in Morocco.

9 Lézine, 1971; Djelloul, 1995; Cressier et alii, 2001; Mahfoudh, 2003.

10 Chabbi, 1967-1968.

11 Louhichi, 1997.

12 E.g. Cressier, Rammah, 2006. The eagerly awaited report of Cressier and Rammah’s excavations at Ṣabra al-Manṣūriyya promises to transform our understanding of the Fatimid site.

13 See for example, Cirelli, 2001 on Leptis Magna; Ferjaoui, Touihri, 2005 on Zama Regia and Althiburos.

14 Cressier, 2012, p. 119.

15 Fenwick, 2013, pp. 11-12.

16 Mattingly, Hitchner, 1995.

17 See Cirelli, 2001; Leone, 2007; E. Rocca and F. Béjaoui this volume for the value of re-examining old excavation reports and archival records.

18 Fenwick, forthcoming; Ead., in press.

19 On bishoprics, see Maier, 1973.

20 See Pringle, 1981 on Byzantine fortifications; Djelloul, 1995 on medieval fortifications.

21 Cambuzat, 1982; Fenwick, forthcoming; cf. Kennedy, 2002; Id. 2011 and in this volume.

22 See Vitelli, 1981, pp. 15-17, pp. 24-39; Gelichi, Milanese, 2002; Stevens, 2016.

23 Chelbi, Paskoff, Trousset, 1995; the ongoing Tunisian-British excavations at Utica have not found any ceramics dating to the 5th-7th centuries, Fentress, personal communication.

24 Fentress this volume; Potter, 1995.

25 The founding of a miṣr (pl. amṣār) usually adjacent to an existing town is an Arabic principle that originated in antiquity with a ḥāḍir («settlement») established outside towns like Aleppo to house Arab auxiliary troops serving the Romans, see Whitcomb, 1994.

26 There is some dispute about the date and location of the first camps founded by the Arab armies, but certainly by 775, the town we now know as Kairouan had been established in some form, see Mahfoudh, Baccouch, Yazidi, 2004.

27 M’Charek, 1999, convincingly identifies this existing settlement as Iubaltianae and dismisses its earlier identification as Qamūniyya. Nothing is known of the municipal history of the site: it may have been an imperial estate centre or a small road station, but bishops are attested at various councils between 397 and 646.

28 Akbar, 1989; Whitcomb, 2007; Kennedy, 2010.

29 Walmsley, 2007, p. 105.

30 Sakly, 2000, p. 68.

31 Ms of Ibn Abi Zayd, Mahfoudh, 2003, p. 75.

32 Talbi, 1976.

33 Northedge, 1994; Whitcomb, 1994; Hillenbrand, 1999.

34 Mahfoudh, 2003, pp. 19-39. Apparently a ditch was cut in 770 during a siege of Kairouan that was so bad that the inhabitants had to eat their dogs and cats.

35 Ibid., p. 28.

36 Id., 2008.

37 Lézine, 1971, pp. 141-154. Bishops attended Councils in 411 and 533: Mesnage, 1913, pp. 164-165.

38 Al-Bakrī, Description de l’Afrique septentrionale, 81-82.

39 Ibnawqal, Configuration de la Terre, 70.

40 Mahfoudh, 2003, pp.173-209.

41 Marçais, 1954, pp. 9-22.

42 See Johns, 1999, pp. 109-110.

43 Leone, 2007.

44 Pringle, 1981. On the shifting ideal of the city in the 6th century, see Saradi 2006.

45 Leone, 2007, pp. 166-279. See also: Mahjoubi, 1979; Février, 1983; Thébert, 1983; Pentz, 1992; Lepelley, 2006; Benabbès, 2007.

46 For bibliography, see Gelichi, Milanese, 2002; Fenwick, 2013; von Rummel, 2016.

47 For a full account of medieval Leptis, see Cirelli, 2001.

48 Fenwick, 2013, pp. 21-23, for a re-analysis of Duval, 1982; Id. 1990, model of a city reduced by the 7th century to a series of small inhabited nuclei consisting of fortified complexes, a church and a production site, and surviving in this fragmented state until at least the 9th century.

49 Duval, Baratte, 1973; Béjaoui, 1996, pp. 38-39.

50 Duval, 1982, p. 625; Id., 1999.

51 Mahjoubi, 1978.

52 Fentress, 1987; Amara, Fentress, 1990; Fentress, 2000; Ead., 2013.

53 Biskra: al-Bakrī, Description de l’Afrique septentrionale, 111; Baghai: Ibnawqal, Configuration de la Terre, 81; Tobna: al-Bakrī, Description de l’Afrique septentrionale, 109. A similar phenomenon may have occurred at Tahudha, see Cambuzat, 1982, p. 203.

54 See Laporte, 2015 for a more detailed overview of the late antique and medieval city, supplemented by Mahfoudh, 2003.

55 See Foucher, 1964, p. 341, though the argument seems to be largely based on the presence of sculptural fragments (capitals, corbels and a lintel with a chrism) reused in the construction of the ribat.

56 Lézine, 1956.

57 Talbi, 1966.

58 Lézine, 1956, pp. 44-45, suggests that the wall follows Byzantine foundations, a suggestion that is plausible — it is similar in construction technique and design to the 6th century walls at Tébessa (Theveste) and the large citadel of Haidra Pringle, 1981, p. 96, pp. 199-200. It should be noted that the sources give the word banà which can mean to reconstruct or build a new.

59 Carton, 1907; Djelloul, 1995.

60 On the mosque see Creswell, 1989, pp. 353ff.; Marçais, 1954, pp. 23ff.

61 Lézine, 1971, p. 104.

62 Baratte et alii, 2014, pp. 221-222.

63 Février, 1965; Fentress at alii, 1991.

64 Béjaoui, 1996.

65 Jones, 1983; E. Rocca and F. Béjaoui this volume.

66 Mahfoudh, 2003.

67 Mahfoudh, Baccouch, Yazidi, 2004, p. 11; Mahfoudh, 2008.

68 Béjaoui, 1996.

69 Mahfoudh, 2003.

70 Jones, 1984; Thébert, 2003.

71 Bel, 1913; Poinssot, 1958; Fentress, Limane, 2010.

72 Al-Bakrī, Description de l’Afrique septentrionale, 26.

73 Pringle, 1981.

74 Cambuzat, 1982.

75 Goodchild, 1967; Jones, 1983.

76 Fenwick, 2013, pp. 26-27.

77 Mahjoubi, 1978.

78 Al-Bakrī, Description de l’Afrique septentrionale, 56-61.

79 Djelloul, 1995; Hassen, 2001.

80 Pringle, 1981.

81 Carton, 1905.

82 Talbi, 1966.

83 According to Ibnaldun, Histoire des Berbères, t. 4, p. 429; also Talbi, 1966, p. 251, under Abū-Ibrāhīm Aḥmad (853-63), over 10,000 forts (ḥiṣn) were built — all in stone and lime.

84 Djelloul, 1995.

85 Ibid., p. 38.

86 Handley, 2004. On Christianity after the Muslim conquest, see Mahjoubi, 1966; Talbi, 1990; Savage, 1997; Valérian, 2011.

87 Duval, 1982, p. 625.

88 Talbi, 1990, p. 319.

89 Church III in Gui, Duval, Caillet, 1992, p. 25.

90 Gauckle, 1907, p. 794.

91 Marçais, 1954, pp. 29-30.

92 Dunbabin, 1978, pp. 9-22.

93 Mila Carthage: Whitehouse, 1983; Haïdra: Baratte, 1996, pp. 151-153.

Table des illustrations

Légende Map 1. — Map of medieval North Africa (6th-9th century).
Crédits Source: Corisande Fenwick, based on Cambuzat, 1982
URL http://books.openedition.org/cvz/docannexe/image/23737/img-1.jpg
Fichier image/jpeg, 84k
Légende Fig. 1. — Great Mosque of Kairouan and the Simaṭ.
Crédits Photography: Corisande Fenwick
URL http://books.openedition.org/cvz/docannexe/image/23737/img-2.jpg
Fichier image/jpeg, 69k
Légende Fig. 2. — Cisterns of Sidi Dahmani, Kairouan.
Crédits Photography: Corisande Fenwick
URL http://books.openedition.org/cvz/docannexe/image/23737/img-3.jpg
Fichier image/jpeg, 72k
Légende Fig. 3. — Plan of Sbeïtla.
Crédits Modified from Duval, 1980, with additions from Bejaoui, 1990 and 1996
URL http://books.openedition.org/cvz/docannexe/image/23737/img-4.jpg
Fichier image/jpeg, 46k
Légende Fig. 4. — Satellite photo of medieval settlement at Belalis Maior.
URL http://books.openedition.org/cvz/docannexe/image/23737/img-5.jpg
Fichier image/jpeg, 94k
Légende Fig. 5. — Plan of Sousse (9th c.).
Crédits Draw: Corisande Fenwick
URL http://books.openedition.org/cvz/docannexe/image/23737/img-6.jpg
Fichier image/jpeg, 37k
Légende Fig. 6. — Sousse showing great mosque, port and walls taken from the ribāṯ.
Crédits Photography: Corisande Fenwick
URL http://books.openedition.org/cvz/docannexe/image/23737/img-7.jpg
Fichier image/jpeg, 61k
Légende Fig. 7. — Medieval baths at Dougga abutting the exterior of the Byzantine fort.
Crédits Photography: Corisande Fenwick
URL http://books.openedition.org/cvz/docannexe/image/23737/img-8.jpg
Fichier image/jpeg, 67k
Légende Fig. 8. — Plan of early medieval Tocra showing location of Byzantine fort.
Crédits Draw: Corisande Fenwick
URL http://books.openedition.org/cvz/docannexe/image/23737/img-9.jpg
Fichier image/jpeg, 46k
Légende Fig. 9. — Ribāṯ of Monastir.
Crédits Photography: Corisande Fenwick
URL http://books.openedition.org/cvz/docannexe/image/23737/img-10.jpg
Fichier image/jpeg, 64k

Auteur

University College London

© Casa de Velázquez, 2018

Conditions d’utilisation : http://www.openedition.org/6540

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search