Version classiqueVersion mobile

Entre civitas y madīna

 | 
Sabine Panzram
, 
Laurent Callegarin

II.2. — Estudio de casos

An Island in TransitionJerba between the Fifth and the Ninth Centuries

Elizabeth Fentress

Texte intégral

  • 1 Fentress, Drine, Holod, 2009. I am grateful to Corisande Fenwick, Renata Holod, Sabine Panzram and (...)
  • 2 Kennedy, 1985; Fenwick, 2013; von Rummel, 2016.

1The INP-University of Pennsylvania survey of the island of Jerba (map 1, p. 242) took place between 1995 and 1999, covering the entire island with a stratified sample of north-south transects and recovering over 400 sites of all periods. The results, at least for the Roman period, were published over 7 years ago1. Much of what I cover here has thus been out for a while. However, it is worth re-examining what we know about Late Roman Jerba in the context of this conference, and in the light of much information that has emerged in the intervening period. The situation of the island in the context of the «polis to medina» debate is rather peculiar, in that it seems to show a radical de-urbanization of the settlement2. However, this can hardly be attributed to the Arab conquest, in that it took place almost a century before. Indeed, it can be argued that the limited productive capacity of the island made city life possible only when it was commercially connected to other areas of the Mediterranean. The city of Meninx/Girba occupies a preponderant place in the Late Roman settlement of the island, and it is here than any consideration of the economy must begin.

Meninx/Girba and urban production

  • 3 I am grateful to Ali Drine for sharing this information: the inscription will be published in Holod(...)
  • 4 Data from the geophysical survey of the INP-Oxford Utica Project.
  • 5 Tissot, 1884.

2Roman Jerba was apparently characterized by a high degree of economic activity and significant prosperity. The principal city in the late Roman period was Meninx, which as a recently discovered inscription shows, was also known as Girba, a name that eventually became identified with the island as a whole (fig. 1, p. 243)3. The town stretches for two kilometers along the southeastern side of the island and extending inland as much as 500 m, giving it a total area of around 80 ha — it was thus similar in size to Utica, whose layout dates to the period in which it was the provincial capital, and one of the largest towns in North Africa4. Early travelers record that it was surrounded by a wall, although no trace of this is evident today5.

  • 6 CIL VIII 11059: Humphrey, 1986, p. 331.
  • 7 Wilson, 2009.
  • 8 AE 1895, 72 = CIL VIII 22785; AE 1998, 1519.
  • 9 AE 1934, 35.
  • 10 Beschaouch, 1985; AE 1986, 711.
  • 11 Aurelius Victor, Epitome, 35.
  • 12 Mesnage, 1912, pp. 55-57.

3Its architecture is extravagant, with massive buildings and an almost vulgar use of coloured marble. In addition to the forum and its splendidly decorated basilica there were at least two large sets of baths, and all the buildings required to keep the populace amused: a theatre, a stadium, an amphitheatre and, probably, a circus, whose presence we can assume from a mosaic showing four named horses6. A plentiful water supply was assured by a series of at least four small aqueducts, one of which ran directly to the port at the southwestern corner of the town7. Few inscriptions survive, however, victims of the stone-poor island’s tendency to burn them for lime: we have thus no certainty of its municipal status. Two public inscriptions remain: a dedication to L. Minicius Natalis, proconsul of Africa in ad 120-121 by the «Meningitani»8, and the erection of statues by a duumvir in the 3rd century9. A 3rd century funerary from Meninx gives evidence for a confraternity of the «Silvaniani», one of the many in North Africa that revolved around the games in the amphitheater10. Trebonius Gallus and Volusianus were declared emperors at their birthplace, «the island of Meninx which is now called Girba»11. Bishops of Girba are known from the African councils between ad 256 and 52512.

Map 1. — The sampling strategy.

Design: Michael Frachetti

  • 13 Fentress, 2009a, p. 77.
  • 14 Drine, 2009, p. 170.
  • 15 Wilson, 2002.
  • 16 Vitruvius, De architectura, 7, 13, 3. Honey could be used to preserve the dye, however.
  • 17 However, the use of abandoned bath buildings for industrial purposes is common in North Africa, see (...)

4The source of the town’s prosperity was its dyeing industry and the commerce derived from it, rather than its agricultural base. At a point not yet established it was linked to the mainland by a causeway, thus making it an obvious place for exchange between the mainland, the island itself and the sea, while its neighbor and perhaps rival, Gigthis on the mainland, was cut off from traffic from the east. Murex production, for which we have evidence in the early second century bc13, clearly grew in importance over the course of the Roman period. Massive dumps of broken shells, as much as 5 m deep, cover earlier houses in almost 10 ha of the town: a short sondage into those heaps indicated that they were still accumulating through the fifth century ad14. This is certainly one of the most striking archaeological testimonies to urban industry ever found in a Roman city15. It will, of course, have involved a large part of the town’s working population, not only those actively involved in creating the dye, but the whole of the chaîne opératoire, from fishing to weaving and dyeing, as well as the merchants (to whom we will return) who would have sold the products of the industry. The presence of dyeworks is certain, in that the dye itself does not survive for long, which would indicate it was used to dye wool and other materials on the spot16. These were either exported directly, or made up into fabrics. In this period we also witness the growth of the artisan quarter, with ever more numerous kilns scattered through the outskirts of the town (fig. 2, p. 244). Kilns producing Keay 25 amphorae succeeded those of the wine amphorae Mau 35, while a number of kilns producing coarsewares also appear. One kiln seems to have produced vaulting tubes, for a waster of one of these was found nearby. Several kilns were found within the area of the stadium baths may suggest that there was an opportunistic use of the heat from the praefurnia, or, at least, some common stockpiling of fuel. On the other hand, it is not impossible that some at least of the baths, with the large covered spaces they provided, may have been given over to artisan activity of some kind17. The town’s massive public buildings were thus surrounded by an outer ring of quarters devoted to a high level of industrial production. In some cases, like that of the baths, they were also penetrated by it.

Fig. 1. — The city of Meninx/Girba in the fifth century.

Design: Michael Frachetti

  • 18 Rabinowitz, 2009, pp. 213-216.
  • 19 Wilson, Tébar Megías, 2008.
  • 20 Bagnall, Várhelyi, 2009, p. 335.

5The distribution of late 4th and 5th centuries pottery (fig. 3a, p. 245) is, if anything, denser than that of the 1st and 2nd centuries. One of the houses excavated in the city, some 350 m southwest of the forum was part of a small insula, and perhaps used as a workshop when built in the late first century ad18. It was given over to murex production from the third century. Evidence for this comes from a stone probably used as an anvil, and a large hearth; as in a similar production in Euhesperides, there were no purpose-built structures like vats that could be securely linked to the production19. This lasted through the beginning of the fifth century, when two rubbish pits were created that contained a large number of ostraca, on one of which the phrase «habet sāniya» (from sanies, the «bloody») refers to murex20.

Fig. 2. — The artisan quarters: kilns on the urban periphery.

Design: Michael Frachetti

  • 21 Plinius, Historia Naturalis, 9, 60, 127. In the 3rd century, Pomponius Porphyrius glosses Horace’s (...)
  • 22 Notitia Dignitatum Occidentis, XI, n. 70; Delmaire, 1989, pp. 679-682.
  • 23 Stirling, 2001, p. 31. For urban production in late Antique African cities Leone, 2007, pp. 220-236

6Crushed murex shells were also a very common aggregate in mortar floors, and in some cases dumps of them were used as makeups to raise floor levels. The negative effects of the murex industry began to emerge during the 4th century, when previously occupied areas started to be covered with the dumps of murex shells. These dumps occur particularly in the western part of the town. Fifth-century Meninx/Girba, still an African Tyre21, seems to have concentrated on the industry, whose murex «slag heaps» and stinking dye factories must have been a far cry from the elegant houses and public buildings of the earlier period. This apparent increase in production might be due to the efforts of the procurator bafii Girbitani Tripolitaniae, known from the Notitia Dignitatum and active from the third century ad when production was brought under imperial control22. The waste products of the industry were no longer centrally collected, and fouled large areas of the city. This invasion of the domestic areas of the town certainly lowered the quality of life of its inhabitants, but there is little doubt that they were there, and that they continued to produce. In the same way, Leptiminus saw kilns and metalworking invading the foreshore and even the center of the site from the 5th century onwards23.

  • 24 Ghalia, 2009.
  • 25 Ibid.

7Possibly one effect of this development of the murex industry was the building of the Christian basilica well outside the city walls24. Over a kilometer away from the southwestern edge of the city, it was probably linked to it by a road, possibly running along the aqueduct that led to the port. It was, however, a large and splendidly- decorated structure, with a hypostyle hall and a baptistery. Tahar Ghalia dates the church itself to the early fifth century, and the hall and baptistery to the sixth century, but, apart from stylistic considerations, there is no direct evidence for this25. A much smaller church is found at the northeast end of the town, almost equally far from the dumps, but beyond a single lintel, certainly not in situ, we have no further traces of Christianity at the site.

8In the sixth century the population of Meninx/Girba declined rapidly. A cistern fill at the building where we identified murex production closes with the first half of the sixth century, while the survey over all produced very little sixth-century pottery, and only 11 sherds of seventh-century pottery (fig. 3b).

9The return of Byzantine government in ad 530 thus seems to have been the beginning of the end of the production of purple on the site. We will return to this decline after discussing the rural settlement.

Fig. 3. — Occupation of Meninx/Girba in the fifth (left), sixth and seventh centuries (right).

Design: Michael Frachetti

Rural settlement

  • 26 Ben Tahar, 2014.
  • 27 Fentress, Drine, Holod, 2009, fig. 12.6.
  • 28 Fentress, 2001.
  • 29 Fentress, Fontana, 2009, p. 195.

10Before turning to the town’s engagement in interprovincial trade it is worth looking at the countryside, its settlement and the nature of its production. Here the picture is not quite so vibrant, although rural settlement through the beginning of the fifth century held up remarkably well (map 2, p. 246). Settlement in the pre-Roman period was very complex, with several small towns, of which the largest was Būrġū, in the centre of the island, surrounded by villas, villages and small farms. There were also a number of fishing ports, of which one, Ghizin, has been recently excavated by Sami Ben Tahar26. The farms, typically, were the first to disappear, and the larger agglomerations of villages and villas seem to have concentrated the labour force in largely nucleated sites. The villas almost certainly concentrated on arboriculture, producing wine and olive oil. These were export crops, attested to by the numerous amphora kilns found throughout the interior of the island27. The wine was most likely passum, or raisin wine, produced on the island until the middle of the last century28. However, as Meninx/Girba took on an ever-greater role in the island the town of Būrġū faded, disappearing altogether by the fourth century ad. A few new villas were constructed, notably the massive structure at Tala, one kilometer north of the city, with a bath building and an impressive cistern complex29. Through the middle empire the agglomeration of the rural population continued, slowing slightly during the Vandal period when only a few farms disappeared, and the number of villas and villages remained stable. Indeed, the Vandal period is archaeologically invisible. However, by the late sixth century the decline in rural settlement speeded up, and only 16 of the 23 villas of the previous period remained, largely concentrated to the southeast in the immediate vicinity of Meninx/Girba (map 3, p. 247). Only one kiln remained in operation. However, seven new villages emerged, one of them on the old site of Būrġū. These were larger than in the past, and suggest a shift in emphasis from directly-managed fundi to less organized groups of workers whose surplus was extracted by taxes. There were still fishing communities, as sites like Sidi Jmur on the west coast and Aghrir on the east show. Ġīzin, to the north, and Qallāla/Haribus on the south coast were certainly occupied at this time. Both sites exhibit some evidence of luxury in the form of marble veneers and tesserae: the pottery workshops of modern Qallāla are full of marble and granite columns from Haribus. The northwest corner of Ġīzin, immediately along the shore, shows evidence for mosaics as well as hydraulic installations.

Map 2. — Rural settlement between 350 and 550.

Design: Michael Frachetti

  • 30 Fentress, 2009b.
  • 31 Aït Kaci, 2009, p. 239.

11Among the new settlements of the sixth century were two small forts on the first inland ridge met when travelling in from the coast, at the sites of Tāla and Ġardāya30. These were built of ashlars on almost exactly the same plan, measuring 50 × 50 RF, or 15 × 15 m (fig. 4, p. 248). They are distinguished only by the buttresses that mark Ġardāya. At their centres were large, double cisterns, built, like all cisterns in Jerba, above ground level. The buildings were occupied on the first floor, reached, presumably, by ladders or wooden steps. The dwellings consisted of rooms around an open courtyard, in which was found the cistern access and a well reaching down to the ground water. A collapsed wall shows that defenders on the roof were at least seven meters above the surrounding landscape. Excavations at Tāla31 revealed a ditch 3 m wide around the outside of the building, running parallel to the walls and at least 2 m deep. The building was dated to the sixth century by an amphora of Samos Cistern type mortared into the masonry, and a sherd of a Gaza amphora of the fifth or sixth century whose mortar traces also suggested it had been built into a wall. The identical plans of the buildings and their strict Roman measurements suggest they were built to a standard plan and had a military function. So, too does their location on the first ridge, dominating roads that approached from the sea. Tala stands on what must have been the line of a road leading inland from the causeway, while Ġardāya is very close to the line of the old road from Meninx to Būrġū. They suggest a plan by the Byzantine governor for an in-depth defense of the productive areas from the mainland tribes. As we will see, at least one of them remained in operation in the medieval period.

Map 3. — Rural settlement between 550 and 700.

Design: Michael Frachetti

The trading community

  • 32 Fentress, Fontana, 2009, p. 199.
  • 33 Ibid.

12The extraordinarily rich collection of pottery recovered from the site shows clearly the extent of Jerba’s trading connections in the fifth century. It is then that a large number of Eastern Mediterranean amphorae arrived on the site, displacing the earlier Italian imports. Quantities of Late Roman 1 amphorae from Cyprus and Syria arrived, along with imports from the Aegean (Late Roman 2 and Samos amphorae) and sporadic imports from Egypt and Palestine32. The fifth century also sees the end of the local production of transport amphorae, suggesting a decline in demand for the passum wines of Jerba, or perhaps a concentration of effort elsewhere. Curiously, the period sees a massive importation of Pantelleria coarsewares, which seem to have enjoyed a boom here as elsewhere in the Western Mediterranean. With them came a few Sicilian and Calabrian amphorae of Keay type 5233.

  • 34 Várhelyi, Bagnall, 2009, p. 335.
  • 35 Fentress, 2011.

13With these visible items of trade we must include the invisible ones; murex dye, as we have already seen, or, more likely, dyed wool and textiles. With them were probably slaves; indeed, one of the ostraca from the site mentions a «mango», or slave trader34. I have argued elsewhere that slaves, particularly children, were imported from Subsaharan Africa by the Garamantes, sold certainly at Leptis Magna: Meninx/Girba might have formed another terminus of this trade, particularly as the security of Leptis Magna collapsed in the fourth century35.

  • 36 I am grateful to Emlyn Dodd for this suggestion.
  • 37 Plinius, Historia Naturalis, 22, 21, 45.

14Who were these traders? The fifth-century ostraca from the murex workshop and the imported amphorae suggest strong connections in the Eastern Mediterranean. There is no doubt that the ostraca are dealing with trade: several are marked with diagonal lines, suggesting cancelled receipts, or payment notes. They generally take the form ‘πρ [monogram]/noun in the genitive/quantity (including measures like sextarii) price/total. The monogram of π and ρ, followed by a noun in the genitive, might suggest simply ‘πρό’ justifying the price paid for the goods then listed. If we bear in mind the context of the ostraca, inside a murex workshop, it certainly seems plausible that the goods they were receiving were integral to the activities of the workshop. One (no. 25) mentions cargoes, ϕορτ [ίoν]. On another (no. 5) goods are described as ‘Τραρίων’, or products of the town of Trarion in the Troad: it seems not impossible, given the position of the little town, in hilly country covered with pine forests, and the fact that the product was measured in sextarii, that it consisted of pitch, used to line amphorae but also possibly smaller containers36. Two (nos. 11 and 23) refer to or ‘Eἰξαλ [α] or ‘Ỉξιαλα’, which the editors take to mean a product of the town on Rhodes called Ixia, but which in fact might more plausibly refer to the bulb of the same name, whose juice Pliny says is used in place of mastic, or chewing gum, and in various medicines37. It might have had something to do with the murex industry, perhaps as a fixative: in both ostraka it is paired with the word ‘óφικιν []’ which could refer to its destination, the workshop in which it was found, or perhaps modify it, to specify the type of ixia extract referred to. One thinks here of the medieval Latin term officinalis specifying the use of a plant in medicine or herbal treatments. The use of this term is limited to the ostraca mentioning Ixia. In the case of these two products, pitch and ixia, but probably also in those ostraca that have not been deciphered, we are looking at the interdependence, and indeed, the complexity, of the murex trade, and its links to the Eastern Mediterranean.

Fig. 4. — The fort at Tāla.

Design: Ali Aït Kaci / Elizabeth Fentress

  • 38 Le Bohec, 1981, pp. 179-189; Stern, 2008, p. 89.
  • 39 Williams, 2013, pp. 251-266 has established the genealogy for one of these families, showing that i (...)
  • 40 Notitia Dignitatum Occidentis, 11, 52; Volpe, 1996, pp. 281-284.
  • 41 Notitia Dignitatum Occidentis, 11, 65. See also Maiuro, 2012, p. 292.
  • 42 For the baphium, Notitia Dignitatum Occidentis, 9, 71; Grelle, 1994, p. 155. On the history of sett (...)
  • 43 Ibid.
  • 44 Numbers 15: 38: «Speak to the children of Israel and you shall say to them that they shall make for (...)
  • 45 Delos: Bruneau, 1969; Gaza: Ovadiah, 1969.
  • 46 Judeich, 1898, p. 156, with a new reading in Ritti, 1993: AE 1994, 1660. For Jewish involvement in (...)

15Most strikingly, of the 23 ostraca where the language can be determined, 21 are in Greek and only two in Latin (and one of these was not found in the same pits). Now, this would fit well with various communities of foreign traders, but only one community in Roman North Africa consistently used Greek, and that is the Jewish community. Particularly in Tripolitania, Greek is the language of Jewish tombstones: indeed, it has been suggested that Jews may have migrated west after the revolts of the second century in Cyrenaica38. In Africa Proconsularis Greek was also used, but to a lesser extent: only 11 of the 47 Jewish tombstones in Carthage are in Greek, rather than Hebrew or Latin. The involvement of Jews with the textile industry is notable elsewhere in the Mediterranean, along with their use of Greek: a good example comes from Venusia, where both a catacomb and a more refined arcosolium tomb record members of the Jewish community in Greek39. Canusium and Venusia were both centers where imperial textile manufactures, or gynacaea, were found from the time of Diocletian onwards40. At Tarentum we find an imperial baphium, along with another Jewish community41. On Minorca, too, the important community of Jews cited in the letter of Severus may have been associated with the imperial baphium, although we cannot be sure on which of the Balearic islands this was found42. Grelle suggested that their arrival in Apulia was linked to the Diocletianic reorganization of these industries, and based on their expertise in dyeing: the same reasoning applies to the community of Minorca and the hypothetical community of Jerba43. Murex dye was intrinsically related to the Jewish faith, in that a thread, the tekhelet, dyed with murex, was required on the corners of a prayer shawl, and the robes of the priests and the curtains of the temple were dyed with it as well44. A small purple-dyeing plant was found on the shore below the synagogue of Delos, and another in the excavations of the synagogue of Gaza45. The involvement of Jews in a guild of purple dyers may be deduced from an inscription from Hierapolis46. Although Bagnall and Várhelyi, in their publication of the Jerban ostraca, specifically refuse to link them with a Jewish community, it seems highly probable that some of both the merchants and the dyers were Jewish. At this point the position of the Christian basilica 1.2 km from the city perhaps takes on another aspect: its builders and bishops may have wanted to distance themselves as much from the community involved with the baphium as from its refuse.

  • 47 Pace, Sergi, Caputo, 1951, p. 314.

16The murex industry alone would have rendered Meninx/Girba an important emporium throughout the time it was active. Fish products and textiles would have been added to the available goods, while it may be that the island was still exporting passum wine (another product of importance to the Mediterranean Jewish community). Slaves from the Sahara, as we have seen, very probably were sold there — and indeed a link to the Saharan trade is provided by a garment dyed with murex from a tomb of a woman from Germa in the Fezzan47. But it is hardly likely that Jerba was self-sufficient in grain, for which they almost certainly depended on imports. With a large amount of the fifth century population concentrated in the emporium, its production and commerce can be seen as the motor that kept the island running.

Towards transition

  • 48 For recent work on the Justinianic plague see the various essays in Little, 2007. For an assessment (...)
  • 49 Stathakopoulos, 2007, p. 102. The inscriptions, published by Duval, Périn, Picard, 1956, pp. 277-28 (...)
  • 50 McCormick, 2007.
  • 51 Fentress, 2009c, p. 221.

17Like the Vandal conquest, the Byzantine reconquest is archaeologically almost invisible, except for the two forts described above. However, although pottery dating is of course rarely as precise as one would like, as we have seen, the middle of the sixth century shows a definite change in the nature and density of settlement, particularly at Meninx/Girba. The workshop excavated at Meninx I was abandoned, as was the house excavated in the same insula, Meninx II. Throughout the city, the distribution of pottery is radically thinner than in was in the previous century, while the tendency for settlement in villages in the interior increases. It is hard not to relate this to the Justinianic plague of 541-54448. Meninx/Girba was, above all, a port, and its trading connections had for the previous century centred on the Eastern Mediterranean. It may have been the gateway from which the plague spread to Sufetula, where in 543 four funerary inscriptions were set up for children who died consecutively49. The plague might not have wiped out the population of the city, so much as driven them into the countryside, to one of the villages that become so obvious in this period50. In any case it seems a more plausible culprit for the premature disappearance of Meninx/Girba than any decline in Byzantine demand for purple, still the imperial colour. Without the income from murex and related trades the population of Meninx could no longer acquire the grain necessary for its subsistence: we must assume that the villages were largely autarchic. Some communal organization, relating to defense and to the payment of taxes must be assumed for these sites. Byzantine fortifications are noticeably lacking at Meninx/Girba: significantly, the two Byzantine fortlets of Tāla and Ġardāya guard the access to the interior of the island, rather than the old town, where any semblance of urban life seems to have disappeared: following the abandonment of the house in Meninx II, the courtyard was reoccupied by the last habitation on that site. This was composed of a sub-rectangular hut with a floor sunk 50 cm below its contemporary ground level (fig. 5, p. 251)51. Around it was built a rough wall using spolia from the earlier building. A tiny hearth and an almost total absence of any material culture besides fragments of hand-made pottery suggest that the occupation was temporary: the wall could indicate the enclosure of a flock, within the courtyard of the earlier building. The lack of any continuity into the medieval period seems to be confirmed by the finds of early medieval pottery on the site, whose position does not echo in any way those of the 7th century ad except insofar as they are rare. Although sporadic occupation certainly took place, Meninx, by the time of the Arab conquest, was already a ghost town.

  • 52 For comparison of site numbers in African Surveys Fentress et alii, 2004.
  • 53 Hitchner, 1987, 1990.

18The drop in settlement is put in perspective by that of rural settlements elsewhere in North Africa. In the Libyan valleys no new sites are found after the year 50052: Jerba’s limited growth in new sites makes it seem positively buoyant by comparison. The Kasserine survey, too, shows a marked decline in rural sites, particularly the small farms, which suggests that there, too, there was a growing nucleation of the population53. The decline in site numbers, from 40 to 28, is thus lower than on the mainland, and some activity is evident in the construction of new villages. However, the international trade which had constituted Jerba’s economic raison d’être had disappeared well before the Arab conquest.

Jerba and the Arabs

  • 54 Kaegi, 2010, p. 180; Id. 2016, pp. 72-74.
  • 55 Holod, Kahlaoui, 2017.
  • 56 Ibid.

19Although Jerba was not touched in the first Arab raid, which ended in the defeat of the exarch and usurper Gregory and his forces at Sufetula in 647, the island did not escape the second wave of raids against coastal settlements in Byzacena carried out in the late 660’s54. In 667/8 a raid from Tripoli led by the Muslim commander Ruwayfi῾ b. Ṯābit seems to have been particularly violent, and would have involved the taking of booty and slaves55. It is significant that the narrative of the conquest refers to the island as a village (qarya)56, for that is very much what our last archaeological glimpse of Roman Jerba resembles.

Fig. 5. — The excavation at Meninx II, from the east. The building in the foreground is an oval hut, slightly sunken into the ground. Behind it, a wall made of spolia.

Photography: Elizabeth Fentress

  • 57 For a summary of the early medieval pottery of Jerba, Holod, Cirelli, 2011, pp. 170-174.
  • 58 Ibid.

20Pottery from the subsequent period provides evidence for the changes that followed the conquest. Although there is some continuity at one of the island’s kilns, and several forms derive from Late Antique pottery, the commonwares of the eighth and ninth centuries are produced in a rather different whitish fabric, with new forms of jars and pitchers57. Small, globular amphorae are also found, possibly used for a continuing wine production58. As elsewhere in North Africa, no imported pottery of this period was recovered, indicating Jerba’s detachment from the surviving commercial circuits of the Mediterranean.

  • 59 Holod, Kahlaoui, 2017.
  • 60 For the earliest references to a Jewish population in the medieval period, and their involvement in (...)

21Settlement changes, and their chronology, are less easy to read, largely because we are rather at a loss when it comes to distinguishing between the eighth and the ninth century pottery. But in fact what is really telling is the fact that by the time we do see a developed settlement pattern, certainly by the ninth century, it is very different from the previous one, although still concentrated in the southwest side of the island, and numerically almost identical. Rather than concentrating the population in agglomerated villages, the sites seem to be clusters of individual houses, separated by perhaps 100 m, which form groups of as much as 6 or 7 units sited on slight rises (fig. 5). We were so startled by the apparent density of this settlement in this area that a further sample was taken in what appeared to be the densest area, and walked very closely (5 to 8 m distance between walkers). Here we found over 20 sites per square kilometer. These were tiny, and rarely articulated beyond a single small unit. Less than a third of them occupy sites with previous Roman occupation. This seems to argue for a significant dislocation of the original population. Renata Holod suggests that it relates to the arrival of Ibadi groups, together with Berbers from the mainland: indeed, one of the forts, Ġardāya, became the seat of the leader of one of these groups59. How many of the original islanders joined this population is completely obscure, but the disjunction is evident, and it is difficult to talk of any continuity, except in terms of the pottery production. The presence of a substantial Jewish population from the 12th century onward might thus be related as much to a new influx of people as to the survival of the group we have proposed worked in the murex industries of Meninx. They would still have been involved in dyeing and textile weaving, although now with new dyes, based on locally-grown indigo rather than murex60.

  • 61 Fenwick, 2013.
  • 62 Delile et alii, 2015.

22The emporial role of Meninx/Girba thus seems to have disappeared along with the population that depended on it: although already badly damaged, probably by the Justinianic plague, settlement seems to have undergone a truly radical change between the seventh and the ninth centuries, perhaps one of the few instances in North Africa where we can see archaeologically the effects of the Arab conquests61. It is in many ways an unusual city, in terms of Roman North Africa, where urban production on this scale is hardly a common phenomenon, and the concentration of economic activity in an emporium without an agriculturally productive hinterland unique — both Carthage and Utica had abundant grain supplies from their territories. Its fate seems to have depended on the disappearance of that very connectedness that we saw in the context of the murex workshop. In the same fashion, Utica too seems to have disappeared prematurely, as its harbor silted up and it lost its access to the Mediterranean62. In this way the cities on the coast may have been more fragile than those in the interior of North Africa, and far more vulnerable to both geological and geopolitical phenomena.

Notes

1 Fentress, Drine, Holod, 2009. I am grateful to Corisande Fenwick, Renata Holod, Sabine Panzram and Andrew Wilson, for their helpful comments on this paper.

2 Kennedy, 1985; Fenwick, 2013; von Rummel, 2016.

3 I am grateful to Ali Drine for sharing this information: the inscription will be published in Holod, Drine, Fentress, forthcoming. For previous discussion of the name of the town see Drine, Fentress, 2009, pp. 174-176; Beschaouch, 1986, p. 542. See also Desanges, 2003, based on the assertion of Ptolemy.

4 Data from the geophysical survey of the INP-Oxford Utica Project.

5 Tissot, 1884.

6 CIL VIII 11059: Humphrey, 1986, p. 331.

7 Wilson, 2009.

8 AE 1895, 72 = CIL VIII 22785; AE 1998, 1519.

9 AE 1934, 35.

10 Beschaouch, 1985; AE 1986, 711.

11 Aurelius Victor, Epitome, 35.

12 Mesnage, 1912, pp. 55-57.

13 Fentress, 2009a, p. 77.

14 Drine, 2009, p. 170.

15 Wilson, 2002.

16 Vitruvius, De architectura, 7, 13, 3. Honey could be used to preserve the dye, however.

17 However, the use of abandoned bath buildings for industrial purposes is common in North Africa, see Stirling, 2001, pp. 69-72, suggests that it was due more to the existence of large, empty, covered spaces.

18 Rabinowitz, 2009, pp. 213-216.

19 Wilson, Tébar Megías, 2008.

20 Bagnall, Várhelyi, 2009, p. 335.

21 Plinius, Historia Naturalis, 9, 60, 127. In the 3rd century, Pomponius Porphyrius glosses Horace’s reference to afro murice with «significat enim purpuram Girbitarum»: Comm in Horati 2, lemma 183.

22 Notitia Dignitatum Occidentis, XI, n. 70; Delmaire, 1989, pp. 679-682.

23 Stirling, 2001, p. 31. For urban production in late Antique African cities Leone, 2007, pp. 220-236.

24 Ghalia, 2009.

25 Ibid.

26 Ben Tahar, 2014.

27 Fentress, Drine, Holod, 2009, fig. 12.6.

28 Fentress, 2001.

29 Fentress, Fontana, 2009, p. 195.

30 Fentress, 2009b.

31 Aït Kaci, 2009, p. 239.

32 Fentress, Fontana, 2009, p. 199.

33 Ibid.

34 Várhelyi, Bagnall, 2009, p. 335.

35 Fentress, 2011.

36 I am grateful to Emlyn Dodd for this suggestion.

37 Plinius, Historia Naturalis, 22, 21, 45.

38 Le Bohec, 1981, pp. 179-189; Stern, 2008, p. 89.

39 Williams, 2013, pp. 251-266 has established the genealogy for one of these families, showing that it stretches to 7 generations between the fifth and the sixth centuries. For the association with the murex industry Grelle, 1994, pp. 145-149, with ample previous bibliography. Note that a similar arcosolium tomb was found just west of Meninx/Girba and has been ascribed by Ferchiou to eastern immigrants: Ferchiou, 1995, pp. 132-133.

40 Notitia Dignitatum Occidentis, 11, 52; Volpe, 1996, pp. 281-284.

41 Notitia Dignitatum Occidentis, 11, 65. See also Maiuro, 2012, p. 292.

42 For the baphium, Notitia Dignitatum Occidentis, 9, 71; Grelle, 1994, p. 155. On the history of settlement on Menorca Panzram, 2013, on the letter of Severus Bradbury, 1996; Panzram, 2012.

43 Ibid.

44 Numbers 15: 38: «Speak to the children of Israel and you shall say to them that they shall make for themselves fringes on the corners of their garments, throughout their generations, and they shall affix a thread of blue (tekhelet) on the fringe of each corner». For the temple veils Josephus, Antiquities of the Jews, 7, 7, where they are described as dyed with the «blood» of the shellfish.

45 Delos: Bruneau, 1969; Gaza: Ovadiah, 1969.

46 Judeich, 1898, p. 156, with a new reading in Ritti, 1993: AE 1994, 1660. For Jewish involvement in the dyeing industry see further Miranda, 1999, pp. 141-144. For the production of fish products, particularly salsamenta, at Meninx/Girba see Drine, Jerray, 2013.

47 Pace, Sergi, Caputo, 1951, p. 314.

48 For recent work on the Justinianic plague see the various essays in Little, 2007. For an assessment of its archaeological visibility Kennedy, 2007, p. 94, who finds a notable decline in urban building in most sites, particularly Hama and Scythopolis. In the countryside, too, new building declines dramatically.

49 Stathakopoulos, 2007, p. 102. The inscriptions, published by Duval, Périn, Picard, 1956, pp. 277-280, do not appear to have been taken up in the various repertories. See also the description of the plague in 549 in Corippus’ Johannidos, 3, 343-389: he significantly asserts that it arrived from the sea, and that it had emptied the cities of Libya (3.362).

50 McCormick, 2007.

51 Fentress, 2009c, p. 221.

52 For comparison of site numbers in African Surveys Fentress et alii, 2004.

53 Hitchner, 1987, 1990.

54 Kaegi, 2010, p. 180; Id. 2016, pp. 72-74.

55 Holod, Kahlaoui, 2017.

56 Ibid.

57 For a summary of the early medieval pottery of Jerba, Holod, Cirelli, 2011, pp. 170-174.

58 Ibid.

59 Holod, Kahlaoui, 2017.

60 For the earliest references to a Jewish population in the medieval period, and their involvement in the dyeing industry, a document from the Cairo Geniza and a letter of Maimonides: Valensi, Udovitch, 1984, p. 8. Stern observes that their claim to an ancient and eastern origin is typical of all North African populations (2008, p. 89). Now see also the Princeton Geniza Project (https://geniza.princeton.edu/pgp/; 20.08.2016) for the medieval records related to dyeing, the textile trade, and Jerba.

61 Fenwick, 2013.

62 Delile et alii, 2015.

Table des illustrations

Légende Map 1. — The sampling strategy.
URL http://books.openedition.org/cvz/docannexe/image/23767/img-1.jpg
Fichier image/jpeg, 55k
Légende Fig. 1. — The city of Meninx/Girba in the fifth century.
URL http://books.openedition.org/cvz/docannexe/image/23767/img-2.jpg
Fichier image/jpeg, 34k
Légende Fig. 2. — The artisan quarters: kilns on the urban periphery.
URL http://books.openedition.org/cvz/docannexe/image/23767/img-3.jpg
Fichier image/jpeg, 31k
Légende Fig. 3. — Occupation of Meninx/Girba in the fifth (left), sixth and seventh centuries (right).
URL http://books.openedition.org/cvz/docannexe/image/23767/img-4.jpg
Fichier image/jpeg, 33k
Légende Map 2. — Rural settlement between 350 and 550.
URL http://books.openedition.org/cvz/docannexe/image/23767/img-5.jpg
Fichier image/jpeg, 71k
Légende Map 3. — Rural settlement between 550 and 700.
URL http://books.openedition.org/cvz/docannexe/image/23767/img-6.jpg
Fichier image/jpeg, 71k
Légende Fig. 4. — The fort at Tāla.
URL http://books.openedition.org/cvz/docannexe/image/23767/img-7.jpg
Fichier image/jpeg, 73k
Légende Fig. 5. — The excavation at Meninx II, from the east. The building in the foreground is an oval hut, slightly sunken into the ground. Behind it, a wall made of spolia.
URL http://books.openedition.org/cvz/docannexe/image/23767/img-8.jpg
Fichier image/jpeg, 80k

Auteur

© Casa de Velázquez, 2018

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

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search