The Bizarre Bazaar: Early Excavations in the Roman East and Problems of Nomenclature

Book synopsis: The sixteenth Theoretical Roman Archaeology Conference was held in Cambridge in March 2006. Sixty papers were given during the two-day conference and covered the breadth and length of the Roman world. The issues of identity, its expression and recognition, were at the forefront of consideration. Sessions also looked at public and private religion, 'Romanisation' from a zooarchaeological perspective, how theoretical archaeology works in the field and the ways in which all of this (and more) is presented to the public. This volume contains a selection of the papers from all of the sessions that ran during the course of the conference.


Introduction
The nomenclature used to describe archaeological phenomena is a fundamental element of the way in which they are understood. Recently, much has been written on the multiple problems with applying ancient terminology to Roman houses (Allison 2004: 11;Winsor Leach 1997). For a site such as Dura-Europos, excavated in the early twentieth century, the problem is not only with the use of ancient terms but also with the contemporary ones. The relationship between the nature of modern imperialism, colonialism, and the varying interpretations relating to the Roman provinces has been shown to be a formative one (Hingley 2000). Similarly, work on the Roman East, both past and present, as with the region in other periods, has been coloured by the modern understanding of the East and its peoples (Said 1979).
This paper focuses on the site of Dura-Europos in modern Syria as a case study. At Dura, it can be shown that the prejudices and assumptions of the excavators heavily influenced their writing. While this is not surprising, the extent to which it coloured the work of later scholars is. These problems are shown to be manifest in both the terminology applied to the archaeological discoveries and the interpretation of the society at the site. After analysing the terms used, alternatives will be proposed, and the new evidence for the previously misunderstood features will be given.

History of the Site and its Excavation
Dura-Europos lies on the middle Euphrates River in eastern Syria. Although there was habitation in the immediate vicinity extending into prehistory, Dura-Europos was a Hellenistic foundation of the Seleucid general Nikanor c. 303 B.C. (Isidore of Charax 1.3-4), probably firstly on and around the citadel of the site (Leriche 1991;. The city was controlled by the Parthians from c. 113 B.C. until A.D. 165, save for an apparently brief Roman occupation under the emperor Trajan c. A.D. 115-117, of which a triumphal arch outside the city is one of the only traces (Baur and Rostovtzeff 1929: 6-7;Baur et al. 1932: 17;1933: 55-68). The city was taken again by Rome in A.D. 165 and attached to the province of Syria. The Roman period at the site ended with the Sasanian capture in c. A.D. 256 (James 1985), after which the site is generally thought to have been completely abandoned (Ammianus Marcellinus 24.1.5).
Dura was rediscovered by an Indian battalion serving under the British in 1920. Wall paintings uncovered by the soldiers at the site roused immediate interest among archaeologists and the public alike. After some hasty initial investigations (Breasted 1922;1924), the site was excavated for two seasons by Cumont with funding from the Académie des Inscriptions et Belles-Lettres in Paris (1926). A further ten seasons of excavation started in 1928, when excavations were conducted by a joint expedition of Yale University and the Académie des Inscriptions et Belles-Lettres under the scientific direction of M. I. Rostovtzeff, and by the end of the Yale-French Academy excavations, almost 40% of the city had been revealed.
After five decades of neglect, the Mission Franco-Syrienne de Doura-Europos began in 1986, led by Pierre Leriche and a series of Syrian co-directors, with the aims of re-examining the extant remains, preserving them, and continuing the study of the site: this expedition has continued annually, up to the time of writing (Leriche 1986;1988;Leriche and Gelin 1997;Leriche et al. 2004). Work herein is based on a combination of material from the extensive archival records on the early excavations held by the Yale University Art Gallery, including information on the artefacts and their contexts that was never published, and reexaminations in the field under the auspices of the current Franco-Syrian expedition.

Problems of Terminology and Interpretation
Terminology is an essential part of scholarly practice. However, once they are established in the literature, certain terms may actually hinder alternative interpretations. This is particularly the case when terms used to describe architectural features, for instance, are taken as shorthand for behaviour or social practice. When culturally-specific terms are applied cross-culturally they carry significant implications and connotations.
Due to the diversity and preservation of material uncovered at Dura, it has been widely used in comparative studies throughout the scholarship of the Roman Empire. The extensive nature of the excavations at Dura have meant that the site has frequently been used as a case study, particularly with regard to its military material, which includes the wealth of parchments and papyri recovered from the site. However, the results and interpretations of the Yale-French academy excavations, while they remain an invaluable resource, are imbued with many assumptions and biases, particularly with regard to the nature and character of the population of the city.

The Agora and Bazaar
One of the most influential examples of such terminology is reflected in the nature of the description of the marketplace in the city. In the Hellenistic period, the area is referred to, perhaps not inappropriately, as the agora, but then when the city was under Parthian rule, the marketplace was termed by the excavators as the bazaar (Rostovtzeff et al. 1944: passim). These terms have then been taken up by other scholars as an example of the supposed nature of transformation under Parthian, barbarian rule. However, the reality of the situation was more complex. Firstly, there is no evidence for the actual term that the city's inhabitants used for this quarter -though there is a reference to a potter's shop in P. Dura 126, (Welles et al. 1959: 396-398) -or that this term changed. Secondly, the nature of the area did not change; it retained its essentially commercial function, endowed primarily with shops, food and drink establishments, and by the Roman period, a brothel. This commercial character is evident until the city's destruction, and as such, describing the entire region simply as the commercial district would be more appropriate (Fig.1). While there was a building-up of the area in the Parthian and Roman periods, there is nothing to indicate a fundamental change in its function, and though there were certainly architectural changes from the Hellenistic to Parthian periods, there was no need for the imposition of the term bazaar that is being used as a quasiethnographic analogy. The problem with the terminology for the marketplace is then twofold: firstly, there is an implied transformation in the region that never really took place, and secondly, there is a direct analogy made between the ancient and then-modern East. Implicit in this is the relationship between the ancient barbaroi that were the Parthians, and the modern ones, as the excavators saw the people of Syria; these people are at once evoked as barbarians and denied any link to the 'glorious' Greek past.

Andron and Diwan
Similar interpretive problems appear in the houses of Dura, where there are elements both in the nomenclature of the rooms and the interpretation of their function that show the attitudes and assumptions of those interpreting the dwellings (Fig. 2). One room type that occurs in many of the houses at Dura was termed the 'diwan' by the original excavators. Diwan or divan is used frequently, with diwan used to describe rooms in 48 houses, and divan used for a further 15, although of those, six are later referred to as diwans, showing that these two terms were used interchangeably. In some of the earlier preliminary reports the terms liwan or iwan are used for this room, with liwan occurring only in the fourth report (Baur et al. 1933: 80, 150, 159).
The diwan is the name given by the Yale excavators at Dura to the large central room off the courtyard. It was usually with a centred monumental entranceway with double-doors, rectangular, and with the entrance on one of its long sides. It was entered via a few steps up from the courtyard, often with benches around its perimeter, faced north, and was flanked by one or two secondary rooms. Aside from the courtyard, this is the only architecturally identifiable room-type that appears in many of the houses in Dura. The room-type is a vital one; first, it is found in most houses, and second it has recurring features, such as the centred monumental doorway and plaster benches. These elements, together with the decoration in the form of mouldings, friezes, and painted murals, all indicate the originally intended use of the space; that they were being used to show in some way the wealth and status of its inhabitants to visitors, and perhaps even hosted formalised entertainment.
Diwan is actually a word of Iranian origin that is generally used to describe Islamic architecture, it denotes a reception hall in a house or palace (the current term used in Islamic architecture for a private reception hall being diwan-i khass) (Hoag 1977: 405;Petersen 1996: 66). Liwan and iwan are similar, and though depending on the period, usually refer to a vaulted room that is open to a courtyard on one side, a fundamental element of Islamic architecture. Use of the Greek term andron, most likely referring to the same room as that called the diwan, is attested in documents from Dura, both from the first century A.D. (P. Dura 19, Welles et al. 1959:104-109;Allara 2002: 43-44, 52), although one refers to a room in a sanctuary (Rostovtzeff 1934: 114-115). The relationship is presumed because of the description of the andron in the text, as a central room with monumental entrance. Saliou discusses the likely function as a reception area of the andron, in relation to its appearance in the text P. Dura 19 (Welles et al. 1959:104-109), but also suggests that it may denote a suite of rooms, and shows that despite its etymology, the room does not relate to use exclusively by men (Saliou 1992: 92-95). Although it is used at the site, the term andron itself is far from unproblematic (Nevett 1999: 17-19, 38;Cahill 2002: 180-193). The term diwan bears absolutely no relationship to the ancient reality. That the excavators chose terms from the architecture of their contemporary orient to describe the ancient houses shows the type of analogy that they were making in their interpretation, which extended from architecture to society and culture.

Figure 2. Idealised house at Dura-Europos, (after H. Pearson, the Yale University Art Gallery Archive).
In an effort to designate these rooms with a name that is more functionally neutral, I would suggest calling the room type simply the 'principal room', a term which was used by Clark Hopkins in one of the preliminary reports (Rostovtzeff et al. 1936: 140), thus avoiding the problematic use of ancient terms for archaeological features (Winsor Leach 1997;Allison 2004: 11-12). The term 'principal room' additionally does not carry the culture-specific connotations potentially prejudicing discussion of the possible use of space.
The importance of recontextualising artefacts has been shown in recent years, particularly in the work of Penelope Allison (2001;2004). Similar work is possible for many of the houses at Dura, using the object registers made in the field during the 1930s excavations. These registers contain contextual information, usually referring at least to the room in which each object was found, for thousands of objects that were never published. While deposition cannot always simply be correlated with the usage of artefacts or room function, some broad patterns can be ascertained from the data. An analysis of objects from the principal rooms strongly suggests that these rooms were being used for more than simply 'receiving'.
The range of activities that apparently occurred in these principal rooms was extensive. There is evidence for industrial activity, attested by the presence of moulds for such things as terracottas. While the determination of the scale of production is difficult, there is also evidence for the use of these rooms for household production. Mortars and grinders for the processing of grain were found, as well as evidence for textile production in the form of loom weights and spindle whorls. While some of this equipment is very portable, some of it is large and would have not been practical to move, at least on a regular basis, and we can assume for such objects that they were used where they were found. So, while it is possible that the use of these rooms changed on a seasonal, or even daily basis (Foxhall 2000), their use certainly can be said to extend beyond simple reception. However, it must be noted that many of the houses of Dura would have been occupied for several generations by the time of the city's demise. What can be reconstructed from the artefacts is only usage patterns during the final phases.

Haremlik and Selamlik
Another problematic term used by the Yale excavators was that used for what they thought were 'women's areas': harem or haremlik (Rostovtzeff et al. 1944: 78, 80, figure 62, passim), which is a Turkish term for the part of the Ottoman house that is not open to the public (Petersen 1996: 108). This accompanied the use of the word selamlik for the men's areas of the house. The notion that the houses of Dura had a haremlik or selamlik also has implications for the nature of the social organisation at Dura; this term, too, should be taken as inappropriate and discarded from use. This is not to reject the very notion of gender segregation in the ancient world (Nevett 1999: passim), but rather the particular analogy with Ottoman segregation, and the positivistic notion that such separation is always archaeologically visible or manifest in architectural forms.
None of the terms used to describe the houses were explained or justified; that the name for the Roman military building called the praetorium was given a lengthy justification (Rostovtzeff 1934: 205-206), and that in the same preliminary report, where the term diwan first occurs prolifically, terms for the houses are not explained, perhaps implies that the houses, and by extension, their occupants, were seen to be less important than their military overlords. However, part of the reason for the terms used may have been that there was in fact no real body of scholarship on ancient houses at the time, while military architecture was a favoured topic. With that stated, Clark Hopkins had trained at Olynthos with David Robinson (Hopkins 1979: 36), and would have been familiar with the terms in use there, which themselves have been used, not without problem, throughout the Greek world (Robinson and Graham 1938;Cahill 2002). Both harem and diwan are terms shaped by the view of the excavators that these were 'oriental' houses, such as they encountered themselves when in the Near East in the early twentieth century, and they named the superficially similar rooms at Dura for functional areas that they recognized in the modern ones, interpretations that were based on architectural evidence alone.
These anachronistic and biased views extend beyond the room type and room function to interpretations of how particular elements of urban infrastructure were used, such as the water provisions. In almost every house there is a subterranean cavity in the courtyard that have been shown to be cisterns in many cases, which were waterproofed inside for the retention of liquid (Allara 2002: 50;Allara and Saliou 1997: 151-152), though other suggestions have been proposed (Hoepfner and Schawandner 1994: 280, 287). The necessity for collecting as much rainwater as possible in such a climate is obvious. However, the cisterns of Dura were initially interpreted as latrines (for one example, see Rostovtzeff 1934: 49), and the excavator's description reveals much: [t]he cesspools in most of the houses were of the most primitive character: a pit in the centre of the court. In larger palatial houses the lavatories are of a more 'modern' character, connected sometimes with private baths (Rostovtzeff, 1938: 49).
The word primitive in the above quote betrays the true nature of the interpretation. These interpretations have heavily influenced the subsequent scholarship: the cisterns are described as cesspools, and the term diwan is used in the work of the most influential authors on Dura (Downey 2003: 32), and on Syrian housing (Balty 1989: 419). More problematically, the inherent social function of the rooms denoted as men's and women's has been accepted.
What these examples show is that the excavators identified what they found with the modern orient that they found themselves in. The views that they held on their contemporary East are then projected onto the ancient East. Obviously this is doubly problematic, not only in that the ancient and modern East are not directly analogous, but also because the excavators, even Rostovtzeff, seemed to have a view of the contemporary East as inferior and 'other', particularly with regard to non-urban populations. Breasted called the Bedouin of the Euphrates 'treacherous and hostile' (Breasted 1924: 1-2) and writes with malice of the Arabs who he presumed had defaced the paintings of the site; the excavators' reports are peppered with such notions (Rostovtzeff et al. 1936: 1;Hopkins 1979: 29, 43), and in their private correspondence they are even more acerbic towards the local population. Clark Hopkins, who would later become the field director of the expedition, in an unpublished letter to Rostovtzeff wrote of the then field director, Pillet, that: His [Pillet's] fixed idea (and perhaps a true one) is that every Oriental is a scoundrel and must be watched . . ." (Unpublished letter in Yale University Art Gallery Archive, dated December 2, 1928, parenthesis in original).
While the Macedonian population of the city, and later, the Roman army, were both considered civilised, and therefore worthy of study, the majority of the population was considered relatively primitive. However distasteful this may now seem, it is not the intent of the current paper to detract from the accomplishments of Breasted, Rostovtzeff, or the other excavators. The region at the time was quite literally a war zone, and while at the site the excavators lived under the protection of the French army, they were in fact under constant physical threat. That all the excavators were willing to travel long journeys over land and sea into foreign and sometimes hostile land, often at their own personal expense, shows that they were led by a genuine enthusiasm for the ancient past. While we need not judge the original excavators of the site, who were operating in a different cultural milieu to our own, we need not accept the value judgements they made on others.

Conclusion
The nature of academic work is that it is a product of its time, as are the people who write it. It is not the purpose of this study to judge the character of those who have studied Dura in the past, or to detract from their accomplishments, without which new studies would not be possible. Rather, what is intended is that it be shown that the nature of the Roman East can be heavily clouded by early modern views on what it meant to be 'Eastern'. What gets lost in that is the variability of people living under Roman rule.
School of Archaeology and Ancient History, University of Leicester