The Archaeology and Looting of Dura-Europos

Dura-Europos was a town on the west bank of the Euphrates River in what is today the Deir ez-Zor Governorate of the Syrian Arab Republic (Baird Reference Baird2018; James Reference James2019; Leriche et al. Reference Leriche, Coqueugniot, de Pontbriand, Chi and Heath2012). Founded in the Hellenistic period around 300 BC, it fell under Roman control sometime around 165 AD. During the third century AD, it functioned as a Roman garrison town until its capture after siege by the Sassanid Persians in AD 256 or soon thereafter and its subsequent abandonment (though with some sporadic small-scale reoccupation). It was excavated between 1922 and 1924 by the French Académie des inscriptions et belles-lettres and again from 1928 to 1937 by Yale University in coordination with the Académie Française. These excavations uncovered rich architectural, textual, iconographic, and artifactual remains, largely from the Roman period of occupation. Of particular importance was the large assemblage of Roman military equipment that dates to the time of the final siege and that is protected by destruction deposits (James Reference James2004). Careful analysis and interpretation of this material and its context has allowed a convincing reconstruction of some episodes of the fighting, including a gruesome account of mining and countermining operations under the town wall (James Reference James2011). Dura-Europos is one of the most important excavated Roman sites, particularly in the eastern empire. In 1986, the Mission Franco-Syrienne de Doura-Europos embarked on conservation work and recommenced investigations at Dura-Europos, but the project was terminated by the outbreak of civil conflict in Syria in 2011. Serious looting was reported in 2013 at Dura-Europos at a time when the region was contested by the Free Syrian Army and Jabhat al-Nusra, and by April 2014, it was estimated from high-resolution satellite imagery that 76% of the area within the town walls had been damaged or destroyed. Outside the town walls in areas of Hellenistic and Roman cemeteries, approximately 3,750 individual looting pits were counted (American Association for the Advancement of Science 2014:5–9). By the end of 2014, Dura-Europos was under the control of the Islamic State of Iraq and Syria (ISIS) and looting was ongoing, but it has diminished since the military defeat of ISIS in 2017 (Brodie Reference Brodie, Hashemi and Shelley2022:23–24).

Roman Arms and Armor

The looting of Dura-Europos has raised questions about the nature of looted archaeological material that might be appearing on the market or entering collections. Coins are an obvious contender. More than 14,000 coins were recovered during the 1928–1938 excavations, present either as individual (stray) finds or in hoards. The “hoards” comprised groups of Roman coins seemingly lost or buried around either AD 253 or AD 256—dates corresponding to Sassanian attacks on the town (Baird Reference Baird2018:126–127; Bellinger Reference Bellinger1949). Given the widespread use of metal detectors reported while the looting was ongoing (Brodie Reference Brodie, Hashemi and Shelley2022:32) and the rich assemblage of Roman military equipment recovered during excavations, surviving elements of Roman arms and armor would seem to comprise another category of material likely to have been extracted from the site and smuggled out of Syria onto the international market.

I know precious little about Roman arms and armor, and to help me understand more, I turned to Thomas Fischer's (Reference Fischer2019) authoritative work, Army of the Roman Emperors. There, I was shocked to learn that for his illustrations, the author had deliberately chosen to publish images of unprovenanced artifacts rather than those acquired through legitimate and scientific archaeological research:

Notice the equation here between privately owned and unprovenanced antiquities, which is not to say that unprovenanced antiquities do not exist in public collections and that there are no provenanced antiquities to be found in private collections. But his description of “previously unknown types” implicitly recognizes that many unprovenanced antiquities in recently assembled private collections have almost certainly been obtained through looting and illicit trade, quite likely, as he suggests, from the formerly communist countries of southeast Europe (Fischer Reference Fischer2019:29) or, more recently, as I am arguing here, from Syria.

At face value, Fischer's statement seems to overturn completely what I had understood to be archaeological orthodoxy with respect to unprovenanced artifacts’ unreliability for research and publication. But as I engaged more closely with the literature pertaining to ancient arms and armor, I realized that, in fact, Fischer is not particularly unusual in his opinion. It also made sense of an episode a few years ago, when a senior professor of archaeology introduced me at a social event as an “extremist” with respect to the illicit trade. At the time, this surprised me, given that I thought my position was in alignment with professionally accepted ethical standards, including those relating to study and publication. But being called out as an extremist seemed to indicate that in mainstream archaeology, the study and publication of unprovenanced artifacts would in fact be a generally accepted practice. (It was not clear to me what the opposite extreme might be. Actively colluding with the acquisition and sale of known looted artifacts, perhaps.) Reading through the literature now, however, I have begun to think that the professor was right and that I am indeed an extremist.

Some of the publications about ancient militaria that I have been reading dismiss opposition (presumably by extremists like me) to the study and publication of unprovenanced artifacts as being politically correct (Graells i Fabregat Reference Graells i Fabregat2024:61) or ideological (Fischer Reference Fischer2019:70). Yet these claims of an ideological opposition to publication overlook a large body of empirical research. I have already mentioned Gill and Chippindale, but there is, in fact, a long-established archaeological literature describing the pitfalls of conducting and publishing research on unprovenanced material. The opposition to the publication of unprovenanced artifacts is not ideological; it is evidence based and—to my mind at least—well founded. That being the case, I believe it is appropriate here to revisit and review the evidence and arguments that have established the detrimental effects on archaeology of research involving unprovenanced artifacts.

Material and Intellectual Consequences

In 1993, David Gill and Christopher Chippindale published their landmark article “Material and Intellectual Consequences of Esteem for Cycladic Figures.” Using Cycladic figurines for a case study, they described the deleterious consequences of their looting and collecting on archaeological research. Cycladic figurines are stylized marble anthropomorphic figurines that were produced and deposited in the Greek Cycladic islands during the Early Bronze Age (EBA). Cycladic figurines first came to scholarly attention outside Greece in the eighteenth century, when they were dismissed as primitive idols or curiosities, but during the early twentieth century, they were revalued by modernist artists as precocious examples of a minimalist aesthetic. One unfortunate result of this newly formulated “esteem” was an increasing demand for the figurines by public and private collectors that was met by the antiquities market and ultimately fed by the widespread looting of EBA archaeological sites in the Cycladic islands during the 1950s and 1960s. By the time of their article, Gill and Chippindale (Reference Gill and Chippindale1993:625) estimated that “Some 85 per cent of the funerary record of the EBA Cyclades may have been lost through unscientific search for figurines.” The ultimate cost of esteem was archaeological destruction.

Gill and Chippindale went on to describe the damaging consequences for scholarship of the archaeological destruction that the uncritical reception of Cycladic figurines as art had caused. Without any reliable evidence of find spot or find context, it is difficult if not impossible to establish such fundamental archaeological facts as chronology, spatial distribution, stylistic variation, or even function. Furthermore, stripped bare of this basic archaeological raiment, the figurines are then open for integration into frameworks of understanding that are completely divorced from any EBA reality—the past as imagined or wished for (Gill and Chippindale Reference Gill and Chippindale1993:658). In the case of the Cycladic figurines, this has entailed their reception as singular works of art produced by “master carvers,” thereby reproducing twentieth-century ideas of artistic genius. Gill and Chippindale later elaborated that for them, the central intellectual consequence is an “unwitting and unthinking conservatism” (Chippindale and Gill Reference Chippindale and Gill2000:504). They state,

Gill and Chippindale were not the first to denounce the causative connection between collecting unprovenanced artifacts and archaeological destruction (see, for example, Burnham Reference Burnham1975:135–152; Coggins Reference Coggins1969; Graepler Reference Graepler1993; Meyer Reference Meyer1974), but their article was the first comprehensive assessment of the harmful “material and intellectual” consequences of archaeological engagement with unprovenanced material.

In 2013, Elizabeth Marlowe followed up Gill and Chippindale with her book Shaky Ground, which extended their critique to Roman sculpture, her own area of specialization. Although much of her study material was not obtained illegally and has been known for centuries, it is still the case that little is known about find spots or find contexts. She estimated, for example, that about 75% of Roman sculptures discussed in introductory textbooks have no reliable information about find circumstances (Marlowe Reference Marlowe2013:2). Marlowe termed objects without find context “ungrounded,” whereas those with context are “grounded” (Reference Marlowe2013:5). The study of ungrounded objects is necessarily limited by their decontextualization and must focus on the object itself and the technicalities of its production, rather than the broader social and cultural contexts within which it was produced, used, and discarded (Marlowe Reference Marlowe2013:39). One outcome of this object focus is that scholarship favors the “visually spectacular” over unusual or less attractive though grounded objects that might possess greater scholarly potential (Marlowe Reference Marlowe2013:43).

Some of Gill and Chippindale's observations and arguments had been prefigured in anthropology by James Clifford (Reference Clifford1988) and Sally Price (Reference Price2001) in their separate discussions of “tribal” or “primitive” art. The tribal art in question comprises non-Western objects that have been misappropriated and, like Cycladic figurines, have been “by turns curiosities, ethnographic specimens, major art creations” (Clifford Reference Clifford1988:189). But not all “primitive” objects can be upgraded to “art.” The selection process favors formally abstract objects that look “modern” and ignores other objects that depart from this aesthetic and that to modern eyes look crude or unartistic. Again, absent context is key. Clifford (Reference Clifford1988:200) noted, “Ignorance of context seems a precondition for artistic appreciation—a tribal piece is detached from one milieu in order to circulate freely in another—an art world of museums, markets, and connoisseurship.”

The earliest discussion I have found of the harmful challenges posed to archaeological scholarship by unprovenanced artifacts was written by Robert Carr Bosanquet and published in the 1904 final report of the British School at Athens (BSA) excavations at the Bronze Age site of Phylakopi on the Cycladic island of Melos (Bosanquet Reference Bosanquet, Bosanquet, Gardner and Hill1904). On account of its precocity, I think it is worth describing at some length. Melos was the source of volcanic obsidian used for tool production from the Paleolithic onward. The excavations at Phylakopi uncovered what was termed an “obsidian workshop” dating to the EBA. In his contribution to the excavation report, Bosanquet described the workshop (Reference Bosanquet, Bosanquet, Gardner and Hill1904:218–220), but he also set out to discuss the chronological and geographical parameters of the ancient trade in Melian obsidian. He wrote, however, “Before discussing the range of trade in obsidian it is necessary to clear the way by discussing the value of certain evidence which has hitherto been accepted without question” (Reference Bosanquet, Bosanquet, Gardner and Hill1904:224). The unquestioned evidence consisted of obsidian artifacts in private collections.

When the putative excavators had first visited Phylakopi in 1896, they discovered obsidian cores and blades scattered around on the surface and a “sort of tunnel” said to have been dug into the site 25 years earlier (so about 1870; Bosanquet Reference Bosanquet, Bosanquet, Gardner and Hill1904:224–225). Bosanquet believed this tunnel had been dug into the obsidian workshop and was the ultimate source of thousands of obsidian artifacts that had been bought by private collectors in Athens during the early 1870s. He was fully aware of the potential problems for research caused by the looters selectively acquiring better-quality pieces and inventing find spots. When discussing a presentation by Christian Blinkenberg describing and categorizing the artifacts present in a collection subsequently purchased for the Ny Carlsberg Glyptotek, Bosanquet (Reference Bosanquet, Bosanquet, Gardner and Hill1904:227) was worried that “we have no guarantee that these proportions represent those of the original deposit, and not a selection made by the finder.” He also reproduced a page from the acquisition register of George Finlay's collection, which had been donated to the BSA in 1899. It recorded the acquisition in 1872 and thereafter of large quantities of obsidian artifacts from an agent whose “statements as to the sources from which he got them were more ingenious than accurate” (Bosanquet Reference Bosanquet, Bosanquet, Gardner and Hill1904:225). As part of his collecting enterprise, and to raise awareness of the interest of prehistoric stone implements, Finlay had prepared an illustrated pamphlet in Greek and in English for distribution to schoolmasters throughout Greece, asking that he be sent anything that might be found (Bosanquet Reference Bosanquet, Bosanquet, Gardner and Hill1904:225). Bosanquet had no doubt about the consequences for the obsidian workshop at Phylakopi:

Consequently, although Bosanquet presented an early description of the limitations of unprovenanced artifacts for archaeological research, he also offered a similarly early recognition that the unwary publication of unprovenanced artifacts can act to stimulate looting.

A Seductive and Troubling Work

Shorn of archaeological context, then, an archaeological artifact has lost much of its research potential—it is ontologically thin. Study and publication of unprovenanced artifacts is ill-advised, because scholarship will be compromised or diminished. But there are further commercial and intellectually enervating consequences of the sometimes intimate collaborations that can develop between collectors and archaeologists. In the first place, the study and publication of an unprovenanced artifact can identify and authenticate it, thereby establishing its rarity or importance and allowing a realistic assessment of its market value and—ultimately—price. The financial worth of publication to private collectors that is realized either through sales or tax-deductible charitable donations has been confirmed through studies of unprovenanced cuneiform tablets in private collections (Brodie Reference Brodie, Almqvist and Belfrage2016:129, Reference Brodie, Brodie, Kersel and Rasmussen2023b:103–105; Gerstenblith Reference Gerstenblith, Hashemi and Shelley2022:83–84). A publication also becomes part of an artifact's provenance or pedigree, and the scholar's name will remain attached to it for future marketing as a sign of acceptability or legitimacy.

“Publication” is not simply a sterile exercise of distanced scholarship; it can involve a prolonged period of contact and collaboration between the archaeologist and the collector. This social aspect of publication was highlighted most vehemently by Ricardo Elia (Reference Elia1993), who characterized it as a “seductive and troubling work,” noting that the relationship of scholar to collector is one of client or guest to patron or host. Personal relations with a collector and possible acceptance of access, hospitality, and perhaps even financial support can introduce indiscernible obligations that can compromise a scholar's research objectivity or integrity (Elia Reference Elia1993; see also Müller-Kessler Reference Müller-Kessler2005:224). Close personal relations might also deter a scholar from publishing or enquiring too closely about provenance so as not to antagonize or offend a friend or patron (Brodie Reference Brodie2020a:98–99, Reference Brodie, Brodie, Kersel and Rasmussen2023b:107). Provenance is treated as something separate—the personal affair of the collector, perhaps, but not something that should be open to scholarly investigation. Consequently, “publication” can be a socially embedded activity, creating what Pierre Bourdieu (Reference Bourdieu and Richardson1986, Reference Bourdieu1993) would have recognized as a transformative nexus of symbolic, cultural, and economic capital. In real-world terms, Christopher Prescott and Josephine Rasmussen (Reference Prescott and Rasmussen2020), quoting Mathew Bogdanos (Reference Bogdanos2005), have called this a “cozy cabal of academics, dealers and collectors”—one that benefits all those involved. The “cabal” provides scholarly endorsement of collecting and, by extension, the illicit trade of unprovenanced and likely looted artifacts. Additional literature on this problem includes Bonnie (Reference Bonnie, Brodie, Kersel and Rasmussen2023), Brodie (Reference Brodie, Green and Mackenzie2009, Reference Brodie2011, Reference Brodie2019), Justnes (Reference Justnes, Brodie, Kersel and Rasmussen2023), Korsvall (Reference Korsvall, Brodie, Kersel and Rasmussen2023), Lundén (Reference Lundén and Karlson2004), Mazza (Reference Mazza, Brodie, Kersel and Rasmussen2023), and references therein.

In light of the recognized synergies between looting, collecting, and publishing, many professional organizations have now adopted policies or guidelines that prohibit the first publication of an unprovenanced artifact unless it can be shown to have been out of its country of origin by a certain date or exported legally after that date (Bouso Reference Bouso2020; Cherry Reference Cherry, Kersel and Rutz2014; Gerstenblith Reference Gerstenblith, Kersel and Rutz2014, Reference Gerstenblith, Brodie, Kersel and Rasmussen2023; Kersel Reference Kersel, Brodie, Kersel and Rasmussen2023). For example, the Archaeological Institute of America (AIA) and the Society for American Archaeology (SAA) have both adopted policies ruling that their journals or monographs should not be used for the first publication of unprovenanced artifacts (Archaeological Institute of America [AIA] 2020; Society for American Archaeology [SAA] 2023). There are, however, exceptions. The AIA (2020) might allow publications that “emphasize the loss of archaeological context or acquisition history.” The SAA (2023) is not so explicit, but it recognizes that there might be “justification for the publication,” with editors or reviewers acting as the ultimate arbiters.

Axel Guttmann and the Aratis Hispano-Chalcidic Helmets

To develop my discussion by way of an example, I want to describe the recent high-profile publications of some Celtiberian military artifacts formerly in the possession of Axel Guttmann and to consider what lessons can be learned from them. Reading about arms and armor, one will quickly come across the name Axel Guttmann. Guttmann was a German businessman who began collecting ancient arms and armor in 1982. By 1993, his collection comprised 1,200 objects, including 174 ancient helmets (Graepler Reference Graepler, Pelagatti and Bell1995:70–71; Junkelmann Reference Junkelmann2002). It is believed that a major part of his collection derived from an “unimaginable” number of burials in southern Italy (Graepler Reference Graepler, Pelagatti and Bell1995:70–71), acquired at a time when the looting of ancient southern Italian cemeteries was at its peak (Elia Reference Elia, Brodie, Doole and Renfrew2001; Graepler Reference Graepler1993). He died in 2001 (Junkelmann Reference Junkelmann2002).

Not everything in Guttmann's collection was from Italy. Sometime in the late 1980s, a cache of at least 20 crushed helmets of uncertain typology (now categorized as and termed Hispano-Chalcidic) was discovered and illegally excavated in the area of the Celtiberian (Iron Age) oppidum of Aratis, located in the modern-day town of Aranda de Moncayo in Zaragoza province, Spain (González Villaescusa and Graells i Fabregat, ed. Reference González Villaescusa and Graells i Fabregat2021; Lorrio Alvarado et al. Reference Lorrio Alvarado, Fabregat, Müller-Karpe, Marugán, Royo Guillén and Munilla2019; Méndez and Montañés Reference Méndez and Montañés2013). The site of Aratis was first occupied sometime during the mid to late sixth century BC and abandoned after its destruction by a Roman army, probably in 74 BC (Fatás Fernández and Romeo Marugán Reference Fatás Fernández, Marugán, Villaescusa and Graells i Fabregat2021a:121–126). The helmets were exported illegally from Spain and sold abroad. Two were sold in the United Kingdom, but the Swiss-based Spanish antiquities dealer Fernando Cunillera sold 17 of the helmets to Guttmann. A final helmet was sold in Barcelona (González Villaescusa and Graells i Fabregat Reference Graells i Fabregat, Villaescusa and Graells i Fabregat2021:30, Figure 8b).

Between 1989 and 1990, before selling the helmets to Guttmann, Cunillera had visited the Römisch-Germanisches Zentralmuseum (RGZM) in Mainz three times for advice about the helmets, hoping to sell them there. He claimed they had been found on a rocky slope at Aranda de Moncayo. The RGZM declined the opportunity to purchase them, and Cunillera subsequently sold them to Guttmann. At the time, the RGZM notified the Deutsches Archäologisches Institut in Madrid about the helmets so that the relevant Spanish authorities could be alerted.

While in Guttmann's possession, the helmets were restored by Hermann Born of the Museum für Vor- und Frühgeschichte, Staatliche Museen zu Berlin (Born Reference Born1993). Born's work in restoring and authenticating the helmets has been criticized for favoring the external appearance of the helmets for the benefit of “art lovers” at the expense of internal features that might be of more archaeological significance (González Villaescusa and Graells i Fabregat Reference Graells i Fabregat, Villaescusa and Graells i Fabregat2021:20) and described as being decisive in persuading Guttmann to buy them (Graells i Fabregat and Müller-Karpe Reference Graells i Fabregat, Müller-Karpe, Villaescusa and Graells i Fabregat2021:131). This highlights, once again, the central role of conservators in sustaining the illicit trade (Brodie Reference Brodie2017; Tubb Reference Tubb1997, Reference Tubb, Prott, Redmond-Cooper and Urice2013).

After Guttmann's death in 2001, his heirs sold off his collection. From 2003 to 2009, 10 of his Aratis helmets were sold through a series of five auctions at the Hermann Historica auction house in Munich, and three more were sold in October 2012 at Christie's London auction house (González Villaescusa and Graells i Fabregat Reference Graells i Fabregat, Villaescusa and Graells i Fabregat2021:31, Figure 8c). The whereabouts of the four remaining helmets are unknown. Efforts on the part of German law enforcement to stop the Hermann Historica sales and recover material were frustrated by the inadequate cooperation of the Spanish authorities.

In 2012, reacting to the high-profile lobbying of several Spanish archaeologists, the Spanish Civil Guard began investigating the looting at Aranda de Moncayo through two operations named Helmet I and Helmet II (Fatás Fernández and Romeo Marugán Reference Fatás Fernández, Marugán, Villaescusa and Graells i Fabregat2021b; Romeo Marugán Reference Romeo Marugán, Oosterman and Yates2021). During 2013, the Civil Guard arrested two individuals and searched six properties, seizing a total of 9,334 archaeological artifacts. In April 2013, the Aragonese government commissioned the preparation of an expert report to support the Helmet operations, with a remit to study the seized material and to conduct an archaeological investigation of Aratis, which took place through 2014 and 2015. In terms of looting, the investigation discovered that a mechanical digger had dug through occupation layers in 1993 to expose a cemetery, which was possibly the source of one of the Guttmann helmets. The apparent find spot of the Hispano-Chalcidic helmet cache was located from the testimony of one of the arrested individuals, but there was no remaining evidence of archaeological context.

Christian Levett bought seven of the Aratis helmets when they were offered for sale at Hermann Historica, and he subsequently placed them on public display at his Mougins Museum of Classical Art (Burns Reference Burns and Merrony2011). In December 2019, convinced by the work of Spanish archaeologists and law enforcement that they had been illegally excavated and exported, he returned them—without compensation—to the ownership and possession of Spain, where they were placed on display at the Museum of Zaragoza. In 2020, he acquired another helmet from an antiquities dealer in Denver that he also donated to Spain (González Villaescusa and Graells i Fabregat Reference Graells i Fabregat, Villaescusa and Graells i Fabregat2021:32, Figure 8d; Levett Reference Levett, Villaescusa and i Fabregat2021:271).

In 2010, while the Guttmann collection was being dispersed through the Hermann Historica auction house, a project was initiated with the encouragement of the RGZM to publish the Aratis helmets. This resulted in an initial publication in 2014 (Graells, Lorrio Alvarado and Quesada Reference Graells, Lorrio Alvarado and Quesada2014), and a second in 2021 (González Villaescusa and Graells i Fabregat, ed. Reference González Villaescusa, Graells i Fabregat, Villaescusa and Graells i Fabregat2021). Available materials included documentation held at the RGZM, the Born (Reference Born1993) publication, the Hermann Historica auction catalogs, some of the helmets themselves (once they were made publicly available at the Mougins Museum), and some other examples already held in Spanish museums. The project sought to chart a “third way” for the study of unprovenanced artifacts, between the polar opposites of uncritical publication or nonpublication (Graells i Fabregat Reference Graells i Fabregat2024:61). An action protocol was defined, aiming to combine archaeological opportunity with social responsibility, whereby the helmets themselves would be studied along with their provenance, with archaeological research proceeding alongside professional and police actions aimed at recovering the helmets for Spain (Graells i Fabregat Reference Graells i Fabregat2024:54, 61). One demonstrably positive material outcome of the publication project was Levett's voluntary decision to return the material he had bought at Hermann Historica to Spain (Levett Reference Levett, Villaescusa and i Fabregat2021:271).

I cannot speak for the AIA, but it seems to me that these publications of the Aratis helmets, with their explicit aim to “emphasize the loss of archaeological context or acquisition history” (AIA 2020), would not contravene its publications policy. Similarly, SAA editors might have considered there was sufficient “justification for the publication.” The publications of the Aratis helmets do not, perhaps, represent the third way that their authors imagine they do. The research is exemplary in its treatment of the archaeological artifacts, although if I could hazard a constructive criticism of the project it would concern provenance (or “acquisition history”). The project set out with the aim of studying the helmets along with their provenance, but the helmets were the primary research interest. The provenance has been published, as I have shown in my account, but it is summary. For my own research, I would like to have learned more about the actors and agencies involved, along with any associated criminality. The person arrested during the Helmet II operation, for example, was convicted of money laundering (Lorrio Alvarado et al. Reference Lorrio Alvarado, Fabregat, Müller-Karpe, Marugán, Royo Guillén and Munilla2019:114), though little more was said about how or what money was laundered.Footnote ¹ In recent years, the use of the antiquities trade for money laundering has been an international policy priority (Brodie and Yates Reference Brodie and Yates2022), and it would have been useful for more evidence of the operation to have been made available. More generally, there is a growing scholarly awareness that there is often more to be learned from the provenance of a decontextualized artifact than there is from the artifact itself. Provenance refracted as “cultural biography” (Kopytoff Reference Kopytoff and Appadurai1986) has become increasingly important for studying the “modern” reception or entanglement of ancient artifacts (Brodie Reference Brodie2020b, Reference Brodie2023a:328–329, and references therein; Kersel Reference Kersel, Hopkins, Costello and Davis2021:261–272; Mazza Reference Mazza2021; Wirth and Rasmussen Reference Wirth, Rasmussen, Brodie, Kersel and Rasmussen2023; Yates and Mackenzie Reference Yates, Mackenzie, Oosterman and Yates2021). Provenance research of that type would probably have fallen outside the Aratis project remit, however, and I might suggest that any similar project in the future would benefit from the inclusion of criminological or sociological expertise.

Synthetic and Corpus Publications of Unprovenanced Artifacts

The disciplinary ground between the Hispano-Chalcidic Aratis helmets and the desire to publish what Fischer characterized emotively as “homeless finds” is poorly mapped out. There is an unresolved tension between the archaeological aspiration to conduct research into unprovenanced material and the recognition that such research could be compromised and even harmful. The EAA publication policy for its European Journal of Archaeology is not much help. It states simply that “Archaeological material knowingly obtained illegally from unprovenanced sources should not be published in the EJA” (European Association of Archaeologists [EAA] 2022a), which seems to imply that most unprovenanced material (which almost by definition has not knowingly been obtained illegally) would be acceptable for publication. The guidance offered by the EAA's Principles for Archaeological Research probably represents a fuller and clearer statement of the EAA's position (EAA 2022b). EAA Principle (2d), “Publication of decontextualised archaeological artefacts,” states the following:

This policy statement seems clear. Unprovenanced artifacts can be of interest for archaeological research and can be published, only if what is known of their provenance is provided. Notice that unlike the AIA and SAA, the EAA does not distinguish between primary and secondary publication. Like the SAA, however, it considers that publication of an unprovenanced artifact might be “justifiable” because it might be useful for archaeological research—provided whatever is known about its provenance is included in the publication—with a view to raising public awareness, aiding law enforcement investigations, or supporting claims for repatriation. Unlike the SAA, however, the EAA does not insist on editorial oversight.

To my mind, EAA Principle (2d) raises at least three issues that require further clarification. First, there is the question of provenance. EAA Principle (2d) neither describes nor offers any guidance as to just what standard of provenance would be acceptable to justify publication. Presumably, it is not simply “find spot unknown” or “private collection.” But would something like “ex Bulgaria, acquired by an English private collector in the 1990s” be acceptable? I suspect it would. Mödlinger and Tsirogiannis (Reference Mödlinger and Tsirogiannis2020) perhaps provides a better example of what is envisaged in her publication of 13 European Bronze Age helmets and a pair of greaves, in which she describes each object in turn and offers a comprehensive account of what is known about their find spot and provenance (Mödlinger and Tsirogiannis Reference Mödlinger and Tsirogiannis2020:329–332).

There are also the inevitable problems of verification. How is it possible to know if what is published about a provenance is factually correct? Mödlinger provides a good example of how a provenance might be invented (Mödlinger and Tsirogiannis Reference Mödlinger and Tsirogiannis2020:330). Her helmet no. 2 was said to have been found most likely in Ukraine. It was sold at violity.com on July 9, 2017, and discussed on a Ukrainian metal detectorist forum on July 13, 2017. After restoration, on September 4, 2018, the helmet was sold at TimeLine Auctions in the United Kingdom, with a provenance stating that it had been acquired in Germany after World War II and thence by descent in 2006 to a lady living in Kent, United Kingdom. The helmet was bought by the Austrian Galerie Kunst der Antike, which sold it shortly thereafter. In 2020, its whereabouts were described as a “private collection” (Mödlinger and Tsirogiannis Reference Mödlinger and Tsirogiannis2020:330). Mödlinger has been assiduous in searching out the provenance of the piece and revealing the deception. But would all archaeologists be so careful? Would simply reproducing the provenance offered by TimeLine or some other commercial organization comprise “all details about the doubtful origin” and be considered acceptable by the EAA? The question is essentially one of due diligence, although the EAA offers no guidance as to what would constitute sufficient due diligence and acceptable provenance.

Mödlinger ends her provenance description of helmet no. 2 by stating that “investigations into suspected offences are ongoing in the United Kingdom and Ukraine” (Mödlinger and Tsirogiannis Reference Mödlinger and Tsirogiannis2020: 330), but she does not reveal how she knows about the investigations and whether or not she was involved (which would be in accordance with EAA Principle [2d]). Perhaps she was maintaining confidentiality, and perhaps for good reasons. I wonder if sometimes an archaeologist might know more about provenance than it seems wise to publish for fear of legal action or personal attacks. I have personal experience with both. It is laughable to read that my position in arguing against publication is considered to be a comfortable one (Graells I Fabregat Reference Graells i Fabregat2024:61). I am sure that it is not as comfortable as enjoying the hospitality of a wealthy private collector while studying unprovenanced objects. Roberta Mazza (Reference Mazza2018) has bravely described how she received threats of acid attacks when researching the provenance of a “remarkable” Coptic papyrus that had appeared on the market. Her article should be required reading for any archaeologist interested in this subject.

Strangely, EAA Principle (2c), “Ethical practice in expert evaluation of archaeological material” (EAA 2022b), does offer strong recommendations about what would be required in terms of provenance and due diligence to ensure ethical practice. It is not clear why these strong recommendations regarding evaluation do not apply to publication. This may be because expert evaluation is regarded as active intervention, whereas publication is not.

The second issue raised by EAA Principle (2d) concerns impact. I am not convinced that academic publications containing even the more detailed provenance descriptions of the type offered by Mödlinger would, as the EAA hopes, “raise public awareness about the loss of archaeological context, the dynamics of the art market and the history of collecting” (EAA 2022b) because I simply do not think the general public (or policymakers) would read them. Collectors, too, would be insulated from them. They surround themselves with sympathetic academics and other experts (the “cabal”) who would be unlikely to draw a collector's attention to anything that is considered harmful to their interests. Having said that, although an academic publication might still act to identify, authenticate, and contribute to the commercial pedigree of an artifact, if it is envisaged that the publication will go unnoticed by collectors and be ignored by the market, does that mean it will exert no discernible commercial effect (Mödlinger Reference Mödlinger2017:12)? Perhaps, but that remains to be determined.

Finally, there is the issue of research quality. EAA Principle (2d) fails to engage with the compromised research potential of unprovenanced artifacts—Gill and Chippindale's material and intellectual consequences and Marlowe's shaky ground. It states that a “decontextualised object can still bear relevant information, and hence might be of interest for archaeological research,” but it does not explore the nature or utility of the “relevant information.” Is the publication of an unprovenanced artifact worthwhile? Is anything learned? From my limited and admittedly inexpert acquaintance with the literature pertaining to ancient arms and armor that accommodates unprovenanced material, I get the impression that it is largely typological, technological, and functional. The value of Guttmann's collection was explicitly described as comprising its importance for understanding the typological, artistic, functional, and manufacturing aspects of the artifacts it contained (Junkelmann Reference Junkelmann2002), confirming Marlowe's (Reference Marlowe2013:39) criticism that research into a decontextualized artifact prioritizes the object itself over the social and cultural contexts of its production, use, and discard. The opportunities for historical reconstruction offered by scientifically excavated assemblages such as those from Dura-Europos have been lost. But even if the constrained research potential of an unprovenanced military artifact is understood, and Marlowe's disapproval accepted, is it nonetheless beneficial to “at least slightly increase the small material framework that survives” (Mödlinger Reference Mödlinger2017:12)?

To be fair, a Roman helmet is less open to fanciful interpretation than a Cycladic figurine (although see Bishop and Coulston's [Reference Bishop and Coulston2006:47–48] discussion of the “Sword of Tiberius”). But publications of unprovenanced artifacts that might be of extraordinary interest because of their beauty, rarity, or technology (Graells i Fabregat Reference Graells i Fabregat2024:60) are still open to Marlowe's criticisms (Reference Marlowe2013:43) about prioritizing the visually spectacular. It is also sobering to remember the observation that Born's conservation of the Aratis Hispano-Chalcidic helmets favored the artistic over the archaeological, as did the conservation of the Roman “Crosby Garrett” helmet found in 2010 by a metal detectorist in the United Kingdom (Worrell et al. Reference Worrell, Boughton, Pearce and Breeze2018:32–33). Important finds might be fraudulently elaborated during “cleaning” or “conservation work” and be archaeologically devalued (Travis and Travis Reference Travis and Travis2014:31)—the danger of “overrestoration and falsity” already highlighted in the quote reproduced above from Chippindale and Gill (Reference Chippindale and Gill2000:505). The unprovenanced artifact received for archaeological study is not always the one that came out of the ground. And, of course, there is the ever-present danger of outright forgery (Bishop and Coulston Reference Bishop and Coulston2006:38; Fischer Reference Fischer2019:70).

I have described the “seductive and troubling work” of the “cabals” of scholars, collectors, and dealers, but it is important to recognize that Fischer and Mödlinger have not published private collections. They have published a scholarly synthesis (Fischer) or a corpus (Mödlinger) of material, each of which includes some artifacts that are of uncertain provenance and find context, some in private collections. Consequently, not all of the arguments against publication marshalled by Elia, me, and others need apply, given that the publications do not involve a close personal relationship with a collector and there is no obvious reputational benefit for a collector that might encourage further collecting. There is no “cabal.” I propose then to distinguish between “collection publications” of the type criticized by Elia and others and what I will characterize for the purposes of this article as “synthetic or corpus publications,” which avoid intimate collaborations with collectors. I think the EAA has synthetic and corpus publications in mind when ruling that the publication of an unprovenanced artifact is justifiable, although I am not sure that is clearly spelled out.

Guilt-Free Publication of Unprovenanced Artifacts?

I have described my reservations about EAA Principle (2d)'s guidance with respect to provenance, impact, and research. To play devil's advocate, it is instructive now to view EAA Principle (2d) and associated statements through Simon Mackenzie's (Reference Mackenzie2014) lens of “conditions for guilt-free consumption in a transnational criminal market,” reframed here as “guilt-free publication.” Should EAA Principle (2d) be viewed as a materially effective solution to the problem of publishing unprovenanced artifacts, or is it more performative in character, in that it provides the appearance but not the reality of “justification,” allowing the guilt-free publication of what might otherwise be proscribed artifacts? Before starting this particular discussion, it is important to note from the word “conditions” that I am not writing about intentional action. I am writing about received conditions of practice—something akin to Bourdieu's (Reference Bourdieu1977) habitus perhaps. Within these conditions, individual actions that might otherwise appear blameworthy are normalized through collective discourse.

Following Mackenzie (Reference Mackenzie2014:505–506), the illicit trade in antiquities can be portrayed as a “public secret,” something that is known about publicly, although with a character that defies definite comprehension. The illicit trade is understood to be “bad,” but the exact nature or severity of its badness remains unknown. Publishing artifacts with an absent, incomplete, or false provenance contributes to the maintenance of this “secret.” It helps to suppress or obscure evidence of criminality and destruction, and it helps the person publishing to avoid any imputations or apprehensions of guilt.

Then there are “language games” (Mackenzie Reference Mackenzie2014:507–508), whereby a sanitizing lexicon is used to avoid describing things in a context of criminality. An example would be “homeless finds” as opposed to “looted artifacts.” The founding document of the EAA Committee on the Illicit Trade in Cultural Material (which does not reference Gill and Chippindale Reference Gill and Chippindale1993) is unfortunate in this respect when it refers to “publicly-made finds” and “‘private’ finds” without defining exactly what these terms mean (Mödlinger et al. Reference Mödlinger, Abungu, Denel, Črešnar, Mele, Özdogan and Thomas2016:16). I despair at the thought of reading about “publicly-made finds from Dura-Europos,” which would provide an accurate though sanitized reference to artifacts looted from Dura-Europos. It would read much better in a published archaeological provenance than the more accurately descriptive “archaeological artifacts looted illegally and destructively from Dura-Europos and traded for the benefit of criminals and armed militia.”

Guilt can also be avoided by blaming the victim (Elia Reference Elia, Nafziger and Nicgorski2009:252–254; Mackenzie Reference Mackenzie2014:507–508). It is widely believed by those publishing ancient arms and armor that the illicit trade is driven by supply: artifacts enter the market from countries suffering from conflict, civil disturbance, or poverty. The flow of antiquities onto the market during the 1990s from countries of the former Yugoslavia and Eastern Bloc in particular has been commented on (Fischer Reference Fischer2019:29; Graells i Fabregat Reference Graells i Fabregat, Villaescusa and Graells i Fabregat2021:252–253; Junkelmann Reference Junkelmann2002). But the idea that the trade is driven by supply goes against most empirical research. Elia (Reference Elia1993:69) maintained that “collectors are the real looters.” Nearly 20 years earlier, Bonnie Burnham had written in less forthright style, “Although dealers and middlemen play an important role in bringing antiquities to the market, it is ultimately the collector who gives the market its impetus” (Reference Burnham1975:135). From a criminological perspective, too, demand is the more significant component of what is understood to be a transnational criminal market (Mackenzie Reference Mackenzie2014:504; Mackenzie et al. Reference Mackenzie, Brodie, Yates and Tsirogiannis2020:50).Footnote ²

Presenting the illicit trade as driven by supply discounts the importance of demand, thereby weakening any imputations of archaeological support for illicit trade through study and publication. But it is, in effect, “othering” the problem (Fabian Reference Fabian1983). It is revealing in this regard to see initiatives aimed at reducing looting and illicit trade in countries of origin inevitably involve training, awareness-raising, and capacity-building (my italics; Brodie et al. Reference Brodie, Kersel, Mackenzie, Sabrine, Smith and Yates2021:2; Lundén Reference Lundén2023:101), aiming to bring “other” foreign professionals and professional institutions up to “modern” standards as they prevail in Europe and North America. These initiatives might be viewed sceptically as concrete exercises in blaming the victim. Countries of origin are seen to have “brought their plight on themselves through poor resource management, bad policing, and restrictive policies which foster underground trade” (Mackenzie Reference Mackenzie2014:507).Footnote ³ Othering the trade in this way abstracts or distances archaeologists from the damaging world of collecting and trade and allows their actions to be presented as disinterested or altruistic—even salvatory.Footnote ⁴

Staffan Lundén (Reference Lundén2023) has examined university curricula provisions as regards illicit trade and has found them wanting. He concluded that there is a need for targeted learning materials in different disciplines. He also expressed concern about what he termed the “hidden curriculum” (Lundén Reference Lundén2023:95–97)—that is, students are socialized through working in a professional environment that normalizes the study and publication of unprovenanced artifacts. For Mackenzie (Reference Mackenzie2014:510), this normalization could be construed as “differential association”—practice learned and accepted within the habitus. Ultimately, it creates the “conditions” for guilt-free publication.

Negative Evidence and Publication Bias

I would propose polemically that the archaeological community in Europe (including the United Kingdom), although not condoning the illicit trade in antiquities, has uncritically adopted a position that creates conditions conducive to the guilt-free publication of unprovenanced artifacts without any reflection on the utility or quality of the published research or its possible consequences. I am sure that not everyone will agree. But the question remains: should synthetic and corpus publications of artifacts that include incomplete provenance information be considered ethical and worthwhile, or should they be viewed simply as exercises in guilt avoidance?

There could be resolution. Alison Wylie (Reference Wylie, Zimmerman, Vitelli and Hollowell-Zimmer2003:10) has characterized the debate over publication as consequentialist. The act of publication could be judged good or bad from the nature of its consequences, with judgement requiring empirical verification of the claims being made about the commercial and scholarly consequences of study. That might sound like a straightforward compromise, but it is not. Most university-based scholars who take an interest in the illicit trade research the trade itself and not its epistemological consequences for archaeological research. There are good reasons why that is the case. To explore fully and expose any vitiating consequences, one would need to be a specialist in the subject area concerned—as Marlowe is with Roman sculpture. But the archaeologists active in a particular subject area do not routinely engage in reflexive consideration of the possible shortcomings of their own scholarly field caused by engagement with unprovenanced artifacts. Marlowe (Reference Marlowe2013:78–79) has made a similar point. The archaeologists who are best positioned to conduct the necessary assessments of harms and benefits have little incentive or inclination to do so.

It might be a problem of “negative evidence” (Lewis and Lewis Reference Lewis and Lewis1980). Disciplinary preconceptions or research paradigms might cause archaeologists to overlook or ignore unwelcome or unpalatable evidence; they certainly do not encourage archaeologists to search it out actively. Even if encountered, intentionally or not, evidence of loss caused by illicit trade that could challenge established practice might not be reported. There are many specific personal and professional reasons why researchers might not report negative evidence, but in general, any research involving an incremental, inductive reasoning process that is conducive to confirmation rather than refutation is more susceptible to this exclusion (Lewis and Lewis Reference Lewis and Lewis1980:555). The construction of artifact typologies has an incremental character, and Chippindale and Gill's (Reference Chippindale and Gill2000:504) caution of “unwitting and unthinking conservatism” comes to mind here. A related problem is that of publication bias or nonpublication (Franco et al. Reference Franco, Malhotra and Simonovits2014). Archaeologists are unwilling to spend time publishing what they regard as negative findings, and peer reviewers are similarly uninterested. The literature with respect to ancient military equipment is populated with positive success stories, but accounts of problematic or contradictory findings that might stand opposed to the success stories are harder to find. Yet, without publication of negative evidence—which might be of little interest to subject specialists—it will not be possible to assess the benefits or otherwise of the publication of unprovenanced artifacts or its ethical standing.

Reflection

The international community is not primarily concerned with the diminution of archaeological research caused by archaeological looting and the illicit trade. It is instead reacting to broader humanitarian concerns, particularly regarding the sovereignty and cultural self-determination of peoples and states and criminal and terrorist involvements. In March 2024, while I was writing this article, the United Nations Educational, Scientific and Cultural Organization (UNESCO) website summarized the problems caused by illicit trade in this way: “Theft, looting and illicit trafficking of cultural property is a crime. It deprives people of their history and culture, it weakens social cohesion in the long term. It fuels organized crime and contributes to the financing of terrorism” (United Nations Educational, Scientific and Cultural Organization 2024). No mention of archaeological research there. Nevertheless, the possibly harmful consequences of studying and publishing unprovenanced artifacts should be a matter of pressing concern for archaeologists because they raise issues that challenge both the material and the ethical status of archaeology as a historical discipline.

In February 2015, the United Nations adopted Security Council Resolution (UNSCR) 2199, which expressed concern about terrorist financing and called on national and international law enforcement organizations to prevent the trade in cultural objects exported illegally from Syria (United Nations 2015). I am starting to believe that military artifacts from Dura-Europos might have been arriving on the market unnoticed or at least without public or scholarly comment for several years now. There is ample evidence that coins from Syria have been traded and sold in Europe and beyond (Brodie Reference Brodie, Finkel, Fraser and Simpson2023c; Reference Brodie, Ulph and Vigneron2025), and I see no reason why the same should not be true of small metal artifacts. In light of UNSCR 2199 and associated policy implementations, however, it seems highly unlikely that traders or intermediaries would include the word “Syria”—not to mention “Dura-Europos”—anywhere near even the sparsest of provenances. Perhaps in time, these artifacts might start to appear in synthetic or corpus publications, qualified as “find spot unknown,” perhaps with a provenance trail leading back to Bulgaria or only to an “English lady.” Many if not most artifacts from Dura-Europos probably did pass through Bulgaria, and a Bulgarian find spot might be convincing enough for an archaeologist willing to publish artifacts from Bulgaria but who might balk at publishing a Syrian one.

An assessment on the ground of the “devastation caused by looting” at Dura-Europos will be the first priority of any future work at the site (James Reference James2019:318). I imagine the same is true for many sites in Bulgaria and adjacent countries where I do not think the damage and associated criminality, violence, and corruption have ever been systematically and comprehensively investigated (although see Center for the Study of Democracy 2007:177–197; Dikov Reference Dikov2019). But the challenge for archaeologists not working on the ground at Dura-Europos or elsewhere in Syria (or in Bulgaria) will be to conduct research into their own disciplinary practices and to reflect on some of the issues I have raised in this article about the potentially recursive relationship between archaeological scholarship and the ongoing looting of archaeological sites. A good start would be to read Gill and Chippindale (Reference Gill and Chippindale1993).