Umm el-Jimal

A late-antique town in its context

-- excerpted from Chapter 14 of Umm el-Jimal: A Frontier Town and its Landscape in Northern Jordan. Volume I. Fieldwork, 1972-1981. Journal of Roman Archaeology. Supplementary Series 26. Portsmouth, RI. 1998.

Bert de Vries

Because of its remarkable preservation, it is tempting to attribute to Umm el-Jimal a unique stature in late antiquity. Viewed in its isolated splendor, it is easy to explain it as a caravan city located at the head of the Wadi Sirhan (Glueck 1942: 3-8; 1951: 1-34; de Vries 1985: 25455). But in antiquity it was but one of many similar rural settlements which fill the S Hauran in a broad band running E-W from Deir el-Kahf to Deraa between the border of Provincia Arabia north of Bostra and Azraq on the south. Some of them, like Umm. el-Quttein, were close to the same size. Once it is realized that Umm el-Jimal was only one town in a densely populated region, it becomes clear that caravans proceeding from Azraq to Bostra had numerous choices for way-stations. The well-developed road network of the S Hauran also included a route from Azraq via Umm el-Quttein (Kennedy, MacAdam, Riley 1986) and via Deir el-Kahf (Bauzou 1985: 138, fig. 1). There is also increasing evidence for an imperially maintained E-W road from Umm el-Jimal via Umm el-Quttein to Deir el-Kahf (Parker 1986: 256-58; Kennedy 1982).

It is clear that Umm el-Jimal was part of a network of settlements in which agriculture was the basic activity, and in which villages and towns were graded by size, according to their regional significance, as market centers for locally-grown agricultural produce. In this scheme, Bostra had the stature of 'central place' (Berry 1967; Hopkins 1980), not simply as the provincial capital and a chartered polis, but as the focus of the regional agricultural marketing grid. Places the size of Umm el-Jimal could have had the r6le of secondary markets, Umm es-Surab tertiary, and so on down to the smallest villages (see note).

The picture of a countryside dotted with prosperous rural communities is not limited to the Hauran. To judge on the basis of the numbers of churches found, Khirbet es-Samra and Rihab were similar in size and prosperity. Though located W of the Hauran, these, presumably like Umm el-Jimal, were in the diocese of Bostra (Piccirillo 1981: 69). The work at Umm er-Rasas (Kastron Mefa'a) is producing a similar picture of a flourishing countryside in central Jordan around Madaba; the Swiss excavators of the 2 churches built into the E section of the fortified enclosure have concluded that the early 6th-c. construction of churches marks an adaptation of a pre-existing fortress for civilian purposes (Bujard 1992: 297). Piccirillo's excavation outside the enclosure (1988:'208-17, 227-31) has demonstrated that ecclesiastical construction was prevalent late in the 6th c. (the Church of Sergius) and continued into the 8th (the Church of St. Stephen). Further work on dating Mefa'a's fortifications and the establishment of the civilian community within them will provide valuable parallels for understanding the contemporary sequences at Umm el-Jimal (see note).

In sum, the countryside of the 6th c. from the E Mediterranean coast to the desert was densely dotted with an array of towns and villages (Cameron 1993: 177-80). Ironically, this social and economic prosperity came at the very time that imperial military controls were weakening, at least in places like Umm el-jimal. While allusions to imperial and military connections at the site are not uncommon from the 2nd-4th c.(see note), they virtually disappear thereafter; the dedicatory inscription of the later castellum (A.D. 412) and a possible fragment of the Anastasius Edict (PES 111: 156, no. 272) late in the 5th c. are exceptions. From the 6th c. there is only a hint of popular involvement in the street politics of the empire in the slogan "Conquer, fortune to the Blues!" (PES III: nos. 256, 262). Alongside it on the walls of the tower of the later castellum other appeals are for protection to God and his angels, not to the emperor and his army. Architecturally too, the military character of Umm el-Jimal gave way to domestic. The later castellum received a thorough remodeling that included the addition of 2 towers, a chapel, and numerous Christian inscriptions. Whether the building continued to have a military function or was turned into a monastic convent cannot yet be determined.

Increasingly, security in the hinterland of the Orient was entrusted to tribal federations, Tanoukh, Salih and Ghassan successively, from the 4th to the 6th c. (Sartre 1982: 132-88; Shahid 1984; 1989; 1995), with formalization of the arrangement in 530 by Justinian. After the peace with Persia in 532, many of the troops garrisoning the frontier of Arabia were discharged and, seemingly, imperial maintenance of fortifications stopped (Parker 1987: 821-22). Bauzou's survey of the N sector of the Via nova has shown that regular imperial maintenance had already ceased in the 4th c. (1985: 143).

All this, combined with the mobile character of the Ghassanid army, indicates that the 6th-c. population of Umm el-Jimal was increasingly marginalized on a fringe of the empire and responsible for its own security (cf. Kaegi 1992: 26-65), especially as wars with Persia farther north preoccupied the emperors and as plagues devastated population centers in the heart of the empire. Fiema has observed a similar pattern of abandoning local population to its own strategic resources in southern Jordan, Palaestina Tertia (Fiema 1992: 329). The reconstruction of the town wall of Umm el-Jimal in the 6th c. (see chapt. 7) and the inclusion of high towers on buildings like the later castellum and House XVII were presumably the work of residents rather than imperial authorities. Such local construction would explain the lack of dedicatory inscriptions by the authorities on structures as substantial and monumental as the tower of the later castellum. The inhabitants needed the look-out towers to observe persons approaching across the fields, especially after 580 when the Ghassanid phylarchate ended (Sartre 1982: 189-98; H. Kennedy 1985a: 166-68). The fence-like and narrow perimeter wall was well within the means of local builders, and sufficient to stop threatening wanderers and retain domestic animals, but neither it nor the residents could have withstood an invading army of Persians or Muslims (see note). The picture of 6th-c. Umm el-Jimal is of a resourceful, self-reliant community that flourished for nearly a century at the very time when the old imperial controls were fading and their Muslim replacement had not yet arrived.

The growth of Umm el-Jimal was part of a vast numerical expansion of substantial settlements in late antiquity. A comparison of ER, LR and (pre-Islamic) Byzantine settlement patterns in the vicinity of Hesban involving 148 survey sites revealed that 57 of them showed evidence of ER occupation, 45 had LIZ, and 126 had Byzantine pottery (lbach 1987: 170-86). Even without qualitative analysis of these sites, it is clear that in the Byzantine period human use of the countryside around Hesban was much more extensive than in the Roman. This area of 100 km2 appears to be representative of the region, pointing to a massive demographic expansion into the countryside in the Byzantine period. The Roman period represented the growth and major build-up of great cities like Gerasa, Philadelphia and Bostra, while the surrounding countryside was settled in a more ephemeral and impermanent manner. At the time when the monumental centers of these major, formally founded cities came to suffer neglect and abandonment (H. Kennedy 1985b: 3-27; Cameron 1993: 168-69), prominent and sturdily constructed settlements were thriving everywhere.

Because strong communities which grew through their own spontaneous efforts are a distinct feature of late antiquity, it is worth comparing them to older, larger cities. The distinctions between 'city', 'town' and 'village' have been much discussed, and the terminology used has not been consistent either in antiquity or today (Haldon 1990: 99-102). For the sake of consistency, this book has restricted the use of the term 'city' to a formally chartered Greek or Roman administrative or commercial foundation (a polis or civitas), major settlements like Bostra, Jerash and Philadelphia. The term 'town' has been reserved for unchartered settlements that nevertheless achieved significant size and autonomy as agrarian and market centers; typical examples are substantial outlying settlements that achieved an apparent measure of informal autonomy in late antiquity, like Umm el-Jimal, Umm el-Quttein, and Shivta. Though such towns may in fact have been as large and wealthy as the older cities, they lacked the formal layout and monumental public buildings of the chartered poleis. Even that distinction is blurred in late antiquity, however, because the older cities gradually lost their monumental character as theaters and temples fell into neglect; shops, houses and churches spilled into the once well-defined colonnaded streets, and the classic meaning of the term polis eroded in the writings of contemporary Byzantine scholars (see Haldon 1990: 101). Finally, we have applied the term 'village' to the 1st-4th c. village of Umm el-Jimal (see note) simply to distinguish it from the town, and because it was about half the size of the later settlement. Formally, such villages were dependencies of the polis in whose territorium they were located, a situation that still prevailed for the hinterland of 4th-c. Antioch, as portrayed in the writings of Libanius and Julian. However, as cities lost control over their territories in late antiquity, such villages acquired greater autonomy, which was eventually encoded in the Byzantine legal system (Haldon 1990: 132-41). Hence, a formal distinction between town and village became vague, reduced to a matter of size and market activity. This is a somewhat arbitrary simplification of a complex set of socio-political structures, but in this study they serve to distinguish three real entities: Bostra, late-antique Umm el-Jimal, and 1st-4th c. Umm el-Jimal. At Unun el-Jimal it will be possible to study the shift in the relationship with Bostra because the earlier village and the later town are separate geographic entities that lend themselves to clearer archaeological comparison than two communities stratigraphically superimposed on the same spot. It is already clear from epigraphic sources that the 1st-4th c. village was closely tied to Bostra. A religious connection is indicated by the altar dedicated to Dushara (PES IV: 34-35, no. 38) as worshipped at Bostra. A political connection is evident in the attribution of Bostra's city council membership to three Umm el-Jimal. residents in their tombstone inscriptions (see note).

The contours of the larger historical framework into which fits the shift in relationship between Bostra and Umm el-Jimal have been sketched by P. R. L. Brown in his study of the redistribution of power by which the church's clergy replaced the urban curia as the notables of late-antique society (Brown 1992). Whereas the curia's basis for power had been exclusive, limited to the citizens of the city, the bishop's basis for power was inclusive, enveloping the larger society of Christians. While it is clear that this included the urban under-class (cf. the bishop as "lover of the poor," Brown 1992: 89-103), it is implied that it incorporated rural Christian society too. The r6le of Umm el-Jimal vis a vis Bostra can no longer be seen so much as one location in the territorium of a particular polis but as a people in a diocese of the universal church. This shift has an architectural component, for the bishop also took over the urban notables' role as sponsors of public building projects (Brown 1992: 120). But whereas the old notables devoted their liturgies chiefly to the beautification and maintenance of the cities, the bishops extended their building activities to rural communities. While the most elaborate ecclesiastical structures are still to be found in Bostra, the plethora of construction in the towns of the diocese represents a vast shift in the distribution of resources previously available for public liturgies from city to countryside (Cameron 1993: 170).

A counterpoint to Umm el-Jimal's relationship with Bostra was its connection with the tribes of the desert. It is our working assumption that the settlers of Umm el-Jimal and the other communities on the basaltic edge of the badiya (see note) are not immigrants from afar, but local Arab tribes, possibly of nomadic background. This assumes that badiya and town co-existed in political, social and economic symbiosis. A suitable regional ethnographic model has been developed by W. and F. Lancaster in their fieldwork in the Burqu' area east of Umm el-Jimal (Lancaster 1988; 1990; Helms 1990: 21-27). The model may more easily be applied to the settlement of the Ist-4th c. village (de Vries 1986: 234-36) because several contemporary Safaitic inscriptions indicate a cultural link with the desert at that time (PES IV: 27881) and the Gadhima inscription implies some association with the Tanoukhid tribal federation in the 3rd c. (Peters 1978: 315-26).

F. Villeneuve has surmised that nomadic tribes interpenetrated the sedentary areas of the Hauran in various ways. He considered a negative relationship - brigandage - the least common, and counted on the positive side sedentarization, sometimes in existing villages; the raising of herd animals, like sheep, goats and camels, in exchange for products like cattle from villages; and the transport by caravan of goods both between locations within the Hauran and to other locations outside (Villeneuve 1985: 117-18). M. C. A. Macdonald concurs with Villeneuve's dismissal of brigandage (1993: 313-14), but cautions that interpenetration during the 1st-3rd c., though one might expect it from analogy with later periods, is not demonstrable from textual evidence. S. T. Parker's thesis (1986; 1987), by contrast, that sedentary communities like Umm el-Jimal needed Roman military defense against nomadic raiders, implying hostile rather than symbiotic relations, has been criticized from several directions (see note).

With minor exceptions, such as the few examples at Umm el-Jimal and Bostra, the 'Safaitic' texts occur only outside sedentary contexts of the steppe and desert (M. Macdonald 1993: 315-16). At Umm el-Jimal the 11 'Safaitic' texts published by Littmann (PES IV: 278-81) are either dedicatory or funerary (not casual graffiti) and appear to be part of the much larger corpus of Greek and Nabataean tombstone inscriptions associated with the 1st-4th c. village. They indicate at minimum that a few preferred to commemorate their dead in the script of the desert, though most preferred Greek. Whether these were visitors, or settled nomads, and whether the majority were settled nomads who switched to Greek and Nabataean, cannot be decided.

That such symbiotic interpenetration occurred in the 6th c. has been assumed by Knauf (1984: 580) who identified the people of Umm el-Jimal as a tribe of the Banu Ghassan. Unfortunately there is only one tribal reference in the contemporary inscriptions at Umm el-Jimal - the Arab inscription from the Double Church, a prayer for Ulaih, the secretary of the chief of the Banu 'Amr (PES IV: 1-2, no. 1). Knauf equated the 'Amr of the text with "'Amr b. Mazin az-Zad, the main unit of the Ghassan" (Knauf 1984: 583). Though it makes sense that there could have been significant associations between the Ghassan and Umm el-Jimal, and though it is difficult to argue the contrary, this identification is tenuous. just as the Gadhima inscription places the teacher of the leader of the Tanoukh at Umm el-Jimal but not necessarily the king and his tribes (Sartre 1982: 134), so this inscription places the secretary of the chief and his tribe there but not necessarily the tribe itself (see note). Other candidates have been put forward by I. Shahid, who would prefer a tribe associated with the Salihids (1989: 324-25).

A strong socio-political connection between the Ghassan and Umm el-Jimal could have been an important factor in the town's relationship to Bostra. M. Sartre has argued that orthodoxy prevailed in the 6th-c. bishopric at Bostra but that the Ghassanid camp had a separate bishop who was monophysite (1985b: 109-18); Shahid, on the other hand, discusses the appointment of a non-resident monophysite bishop to the see of Bostra in the 6th c. (1995: 768-69). The question then is whether Umm el-Jimal was under the ecclesiastical-political authority of the bishop at the Ghassanid camp, or more directly in the sphere of Bostra, as Khirbet es-Samra and Rihab appear to have been, to judge from their churches' dedicatory inscriptions.

NOTES

The agricultural nature of the economy of Umm el-Jimal was a focus of study in 1984. That the S Hauran was part of a larger agrarian network encompassing the entire run-off area of Jebel Druze is becoming clear from the important study of village and field systems done by F. Villeneuve in S Syria (1985).

Back to Section

More distant networks of thriving rural and hinterland communities include those of the Negev (Negev 1977: 647-60, 1980: 3-22; Evenari et al. 1982; Segal 1985; Shereshevski 1991: 20-102), the Upper Galilee (Meyers 1982), central Syria (de Vogue 1865) and N Syria, the region made prominent by the survey of Tchalenko (1953-58).

Back to Section

That these fortifications were designed to defend a frontier (limes) is now questionable as a result of the critical scrutinies of Graf (1989) and Isaac (1992). See further below.

Back to Section

There is no evidence of a similar enclosure wall around the earlier 1st-4th c. village. Such self-protection was not needed in the centuries before the Palmyrene rebellion, when Rome governed efficiently and protected its province of Arabia.

Back to Section

Cf. Villeneuve 1985: 75-76 for a discussion of villages in the N Hauran.

Back to Section

The identification is based on Sartre (agreeing with Littmann) that the abbreviated title "BB" following the names of these individuals is to be read as ????. These inscriptions are ascribed to residents of the earlier village because they were in secondary use in the late-antique town, and are typical of the whole corpus of dated inscriptions from the 2nd-3rd c. (Sartre 1985a; 1985b: 78-79; PES III: 161-62, no. 284; 177, nos. 343-44).

Back to Section

As defined by Helms (1990: 27-30), badiya is the sphere of Bedouin cultural influence that inheres in and stems from the steppe, but as its use evolved in early Arabic sources, political, ideological and cultural aspects became more significant that a strictly geographic one. "The concept develops from a consciousness or an outlook peculiar to tribal social structure in Arabia, and its roots lie in nomadic pastoralism" (29).

Back to Section

Banning's (1986) argument for a symbiotic mutualism rather than hostility in sedentist-nomad relations around the fortress of el-Lejjun is applicable here. Those analyzing the classical sources which Parker cites as evidence for a nomad threat have argued that this textual data is flimsy (Graf 1989) and that it comes mostly from Roman authors guilty of negative stereo-typing (Kennedy 1992: 485). M. Macdonald's study of the nomadic evidence -- "Safaitic" graffiti -- reveals a situation of stable transhumance with regular seasonal movements between the basaltic Harra and the deeper desert to the south and east, not the sporadic raiding of tribes driven from Arabia by adversity (1993: 323-34).