Rereading Pirenne

At the beginning of the twenty-first century, as Europeans and North Americans vigorously debate the place of Islam in the West, Henri Pirenne’s Mahomet et Charlemagne (1937) and his explanation of the clash of civilizations that led to the end of the Roman Empire has gained renewed popular relevance. Although Pirenne’s interest was not in Islam, but rather the contribution of the Arab conquest to the Carolingian renaissance, his work has not only continued to garner attention in the last several decades but, in a disturbing albeit unsurprising development, it has also gained a new set of devotees. Indeed, Mahomet et Charlemagne has attracted readers from among adherents of European right-wing nationalism and American neoconservatism. In their eagerness to condemn the rise of Islamic fundamentalism as a menace to the global hegemony of the West, these historical revisionists have used decontextualized passages of Pirenne’s monograph to argue that he labeled Islam the ultimate danger to Western civilization.¹

This ideological reading of Mahomet et Charlemagne contrasts starkly with the work’s more nuanced interpretation and reception in European medieval and Mediterranean historiography, where what is now known as the Pirenne Thesis has made its largest contribution in economic history.² Although the influence of the monograph’s argument has ebbed and flowed since the time of its composition in response to contemporary trends in medieval studies,³ nearly eighty years after its publication, medievalists continue to debate the merits of Pirenne’s contribution.⁴ Its influence in medieval studies has also varied regionally. Whereas Anglo-American, French, and Italian scholars, on the whole, have generally accepted the research priorities established by Mahomet et Charlemagne while they have questioned its conclusions,⁵ the impact of Pirenne’s work in Austrian and German circles has been markedly less profound than the legacy of his contemporary, the Austrian historian Alfons Dopsch.⁶ Medievalists specializing in the Carolingian period and the epochs that succeeded it or those who work on Mediterranean history more generally have been most apt to embrace Pirenne.⁷ In particular, Mahomet et Charlemagne influenced the future development of the study of Carolingian economics by Pirenne’s successors, like Adriaan Verhulst and Jean-Pierre Devroey, among others.⁸ By contrast, medievalists whose interests focus on the Merovingian period have found Pirenne’s narrative less useful. He looked down on the Merovingian epoch as a decadent period that followed the Germanic invasions and dismissed it as offering minimal innovation,⁹ a characterization that continues to dog the era popularly known (or dismissed) as the Dark Ages.¹⁰

It is striking, then, that a monograph as formative as Mahomet et Charlemagne almost never came to be. The published work that we read today is not a direct transcription of the manuscript found on Pirenne’s desk after he succumbed prematurely to illness in 1935, shortly after the passing of his oldest son Henri-Édouard.¹¹ Indeed, the Belgian historian’s wife Jenny Pirenne (née Verhaegen) and sole surviving son Jacques invited Charles Vercauteren, one of Pirenne’s students, to help prepare a posthumous volume from the unfinished work on which the historian had been laboring up until the time of his demise.¹² Published in 1937, the slim volume honed ideas the Walloon scholar had developed during the First World War but which took more concrete form in articles he composed in the early 1920s.¹³

It appears that Mahomet et Charlemagne derived foremost from Pirenne’s desire to understand the origins of late-medieval cities, the main focus of his research, which Louis Ganshof attributed to Pirenne’s upbringing as the son of a manufacturer.¹⁴ To this end, Pirenne argued that there was a clean break between the cities of Roman antiquity, the lifeblood of which were the Syrian merchants who served as the mediators of long-distance Mediterranean trade in luxury objects, and urban centers that emerged during the central Middle Ages.¹⁵ In Pirenne’s estimation, classical cities died out during the course of the Merovingian period and bequeathed nothing but their names and walls to the new towns and churches that emerged in the late tenth and eleventh centuries.¹⁶

Pirenne also wrote Mahomet et Charlemagne at least in part as a response to the publications of Dopsch.¹⁷ The economic historian, like Pirenne, argued for continuity in the post-Roman period but pointed to a far greater role for Germanic contributions than the Belgian scholar was willing to countenance.¹⁸ Like the French historians Fustel de Coulanges and Ferdinand Lot,¹⁹ as well as the majority of French archaeologists of his day who viewed themselves as heirs of Rome rather than Germania (and are thus known as Romanists),²⁰ Pirenne’s Mahomet et Charlemagne downplayed the impact of the Germanic invasions on European history. While liberally employing reference to Romania for the regions conquered by Rome throughout the text, he avoided the term Germania whenever possible, no doubt because it suggested an independent civilizational structure.²¹ In this way, Pirenne suggested that Roman society maintained its Mediterranean character despite the arrival of what he concluded was an “infime minorité” of barbarian immigrants,²² a view embraced by few scholars in Germany or Austria either in the twentieth century or today.²³

As is often recounted, another important source of influence on Pirenne’s view of the postimperial Roman world presented in Mahomet et Charlemagne was his training in German historical methodology and his cooperation on a number of publications with the German historian Karl Lamprecht in the 1890s and the inaugural decade of the twentieth century.²⁴ However, the death of Pirenne’s son Pierre in the Battle of Yser in October 1914, Pirenne’s internment for civil disobedience in a German prisoner-of-war camp from 1915, and his disappointment with Lamprecht’s (and many other German colleagues’) unflinching support for German imperialism made the Belgian historian more critical of German historiography even if it did not undo the profound impact of German scholarship on his thinking.²⁵ Although Pirenne’s reaction to these events was not as xenophobic as that of his contemporary Emile Mâle,²⁶ these formative experiences lowered considerably his estimation of the German nation, as he noted in a reference to his captivity in a 1920 article in the Revue des deux mondes.²⁷

As has been ably documented by Patrick Geary, Ian Wood, and others, there is no doubt that Pirenne’s composition betrays many of the concerns of the period in which it was written. However, although scholars have given substantial attention to the foundational contributions of national and pan-European perspectives to historical conceptions of the early Middle Ages,²⁸ one perspective that has not been explored sufficiently up till now is that of colonial and postcolonial historiography and the ways in which Orientalism affected nineteenth- and twentieth-century perceptions of the birth of the European Middle Ages. Although Edward Said observed as early as 1979 that Islam represented for Pirenne the Other against which Europe was defined,²⁹ the suggestion has not opened the way to further exploration of this topic. This lacuna is an invitation to reread the Pirenne Thesis not just in its nationalist but also its imperial context and focus as much on its silences as its claims about empires.³⁰ By reengaging the text from this perspective, it becomes evident that some of the assumptions about Islam built into Pirenne’s elegant framework are more vexing than the historical problems he sought to resolve related to culture and economy in the post-Roman world. With the benefit of the wealth of archaeological data that has arisen largely since Pirenne’s composition of Mahomet et Charlemagne, it is incumbent upon the many scholars whose research priorities are still defined by or against Pirenne’s underlying premises to critique more definitively the assumptions embedded in and limitations imposed by Pirenne’s portrayal of the post-Roman epoch. While no interpretation will be able to solve all the enigmas of this period, it is time to relinquish a model too broad in its strokes and problematic in its implications. The transformation of the Roman world was far more variable and complex than Pirenne envisioned at the start of the twentieth century and thus his model should not continue to drive and shape research on this era.

Pirenne and Muhammad

In 1974, Peter Brown praised the work of Pirenne as freeing research on late antiquity from the impasse of the Romanist-Germanist debate,³¹ a historical discussion that divided scholars since as early as the sixteenth century between those who saw the roots of medieval Europe in ancient Rome (Romanists) and those who gave credit to Germanic invasion and migration for the creation of medieval institutions (Germanists). Whereas eighteenth-century Germanist scholars, like Henri, comte de Boulainvilliers, and his successors, emphasized significant discontinuities and collapse brought on by the so-called barbarian invasions, Romanist scholars, like Jean-Baptiste Du Bos and his followers, argued that Roman institutions remained relatively intact through the period of the successor kingdoms.³² However, Pirenne’s work clearly fell in the latter camp, denying Edward Gibbon’s popular narrative of the fissure brought on by the barbarian attacks, and building instead on the foundations established by his Romanist predecessors, like Numa-Denis Fustel de Coulanges in La cité antique (1864). As early as 1939, the Swedish historian and numismatist Sture Bolin characterized the aim of Pirenne’s thesis as reducing the post-Roman era to an extension of the imperial epoch and thus not distinctive in its own right. In an English version of his 1939 article published in 1952, Bolin noted that Pirenne’s work differed from that of Dopsch, since he placed much greater emphasis than the latter on Carolingian innovation.³³ Indeed, Pirenne downplayed the unique (and especially Germanic) features of the Merovingian period in an effort to show its continuity with late-imperial Rome.³⁴ According to Pirenne and many French historians of his era, the medieval period began with the Carolingian period. Its institutions, like feudalism, resembled those of the late Middle Ages rather than those of antiquity, and its apparent contributions overshadowed any contributions by the Merovingians.³⁵ Pirenne’s work was inspiration for the early Annalistes, most notably Lucien Febvre and Marc Bloch, whom he mentored before and after they inaugurated their new journal the Annales in 1929.³⁶

What Brown and others have appreciated about Pirenne’s work was the facility with which it shifted emphasis from the centuries-old Roman-barbarian debate to a discussion of decline and transformation through economic markers such as trade and other quantifiable forms of Roman vitality.³⁷ However, this high praise deemphasized the second and arguably more original part of Mahomet et Charlemagne, wherein Pirenne attributed disruption of Mediterranean unity, which he deemed the most essential characteristic of the Roman Empire,³⁸ to the Islamic conquests of the seventh and eighth century. Pirenne claimed that Arab control of much of the Mediterranean was the most essential development in European history since the Punic Wars, severing East from West and creating a Greek (rather than Roman) Empire that could do little more than defend its fleet and its last possessions, including posts in Naples, Venice, Gaeta, and Amalfi.³⁹ And, although Pirenne’s claim of an Arab-controlled Mediterranean largely devoid of commercial activity was hotly contested by scholars more familiar with Islamic history as early as the 1940s,⁴⁰ the focus of their successive critiques was almost uniquely this economic argument rather than the larger civilizational claims staked by Pirenne. Notably, Maurice Lombard demonstrated that Arab leaders, just as their Byzantine and Persian predecessors, engaged extensively in trade during the period questioned by Pirenne.⁴¹ Bolin, too, pointed to the wealth of data for continued economic interaction even if it was not conducted via the Mediterranean. He forcefully demonstrated that the sphere of trade that included Western Europe was, if anything, enlarged by Islam, which shifted its central focus first to Syria and then to Iraq.⁴²

Moreover, a related and essential dimension of Mahomet et Charlemagne was its discussion of the religious dynamics of the post-Roman West. In his study, Pirenne gave significant attention to Christian institutions, suggesting that they were integral to guaranteeing cultural continuity in the West (just as he believed that disaccord over the definition of orthodoxy in the East ensured the rapid fall of the regions of Syria, Palestine, and Egypt to Islamic domination).⁴³ Importantly, Pirenne viewed the Arab conquest as dramatically different from those of the pagan and Arian Germanic peoples, as it brought about complete reorientation of the lands they controlled:

As characterized by Paolo Delogu, Pirenne’s work effectively captured the transition from the empire to the post-Roman states by pointing not just to changed conditions of political and economic activity but also to altered psychological and religious attitudes. According to Pirenne, whereas leaders of the invading Germanic cohorts might be prepared to acknowledge the inferiority of their culture as opposed to that of Rome, the Islamic invasions succeeded due to the zeal of the Arab conquerors who intentionally subverted the preexisting order.⁴⁶ Whereas the pagan Germanic peoples blended into the Roman Christian population with time, the religious inclinations of Arabs allegedly made this impossible. And, in regions conquered by the Arabs, the heirs of Rome were forced to assimilate even if they were allowed to retain their religious practices: “Germanic peoples were Romanized once they entered into Romania. Romans, by contrast, were Arabized once they were conquered by Islam. It matters little that, until the midst of the Middle Ages, there subsisted in the midst of the Muslims small communities of Copts, Nestorians, and above all Jews. This entire ambience was no less profoundly transformed. There was a break, a complete rupture with the past.”⁴⁷ With this argument in favor of historical discontinuity brought on by Arab conquest, Pirenne noted that the sea that had once been the cradle of Christian civilization now became its frontier.

In Pirenne’s view, the siege of Constantinople loomed large. Following the period of anarchy that succeeded the death of Justinian II in 711, Pirenne noted that Leo III the Isaurian was able to repel the Umayyad fleet definitively in 718 thanks to his use of Greek fire and his alliance with the Bulgarians. Pirenne described this event as “the last attack attempted by the Arabs against the city ‘protected by God.’”⁴⁸ He contrasted Byzantine success with the battle of Poitiers (732), which he saw as less significant than was traditionally held. The conflict lacked the momentous hallmarks of Constantinople’s victory because of the persistence of the Muslim threat in the south of Gaul following the victory of Charles Martel and Duke Odo (Eudes) of Aquitaine over Emir ‘Abd al-Rahman I.⁴⁹ For Pirenne, the Muslim threat remained considerable for Western Europe well into the period that saw the rise of the Carolingians. Pirenne thus considered the Arab invasions to have brought not just economic disruption but also to have altered the very equilibrium of European civilization.⁵⁰

The reasons for this radical stance are not entirely clear. Pirenne’s biography does not suggest that the motivation for this contention was either his religious faith or political conservatism. While inspired by Godefroid Kurth’s Les origines de la civilisation médiévale (1886), Pirenne, a liberal Catholic whose father was a Protestant and whose mother was a devout Catholic, critiqued his mentor’s view that the church was the source of modern civilization.⁵¹ Nor does this perspective appear to have derived from Belgium’s deep entrenchment in and violent colonization of the Congo from the 1880s,⁵² reference to which can be found nowhere in Pirenne’s composition of Mahomet et Charlemagne. Likely, he considered sub-Saharan Africa irrelevant to a discussion of Roman civilization in North Africa.

Although research was still not well advanced in either the history or archaeology of Byzantine or Muslim rule in the Levant or North Africa in the early decades of the twentieth century, some foundations had already been established. Judging from the footnotes of Mahomet et Charlemagne, which may not be entirely reflective of Pirenne’s reading on the subject since these were completed posthumously, he was not particularly well informed about existing scholarship on the Umayyad and Abbasid Caliphates. Pirenne’s characterization of the Arab expansion primarily as a destructive series of razzias seems to have derived instead from stereotypes of Arab culture and society that were prevalent in his day.⁵³ Indeed, Pirenne was not entirely ignorant of North Africa: he traveled to Algeria in 1931, just a year after the centennial celebrations of the conquest of the region, to give a series of lectures attended by, among others, none other than Fernand Braudel.⁵⁴ He likewise made a two-month visit to Cairo in 1934.⁵⁵ Although these visits likely came too late to have shaped his conception of the clash of civilizations in Mahomet et Charlemagne in any profound way, Pirenne’s off-the-cuff depiction of the unwavering devotion of Arab attackers and their rigorous adherence to their faith in the course of their military undertaking reflected an uncritical adherence to contemporary Orientalist stereotypes and colonial discourse that permeated academic undertakings and cultural activities in this period.⁵⁶

Indeed, Pirenne’s basic pronouncements on Arab society mirrored the dominant ideological parlance of French historians of his time with respect to a unified (or Latin) Mediterranean after more than a century of colonial rule and indigenous resistance in North Africa.⁵⁷ His contention that the rise of Islam brought definitive closure of the classical period reflected his contemporaries’ arrogant confidence in Europe’s civilizing mission vis-à-vis its North African and Middle Eastern colonies both in ancient and modern times.⁵⁸ In direct opposition to the manner in which he believed that Catholicism steadily civilized Germanic populations in post-Roman states in the West as they adapted to Roman ways, Pirenne essentialized Islam as a rigid theocracy that brokered no compromise.⁵⁹ Perhaps it was only the upbeat conclusion of Mahomet et Charlemagne that assuaged concerns about the lessons that his contemporaries took from their study of the end of ancient Roman rule in North Africa and the dire consequences they believed would result in the West from a reversal of the modern political order.⁶⁰

More specifically regarding this last point: for Pirenne, the silver lining in the new balance of power wrought by Islam in the Mediterranean was that its closure caused the Carolingians in northwestern Europe to turn inward and innovate politically, culturally, and artistically in what we today know as the Carolingian renaissance.⁶¹ Although their gains were tempered by Viking incursions, this remained the period that Pirenne and his contemporaries linked with the end of Merovingian barbarity and the start of medieval Christian Europe.⁶² For, in contrast to recent neoconservative revisionist interpretations of Pirenne, Mahomet et Charlemagne ended on a positive note. Despite the flaws that attracted challenges, first from historians and numismatists,⁶³ and, in more recent years, archaeologists,⁶⁴ the volume enjoyed long-term success among medievalists precisely because of the manner in which it neatly captured the changes that divided the end of what his contemporaries thought of as the ancient world from the beginning of what we think of as the medieval one. In North America, for instance, Pirenne’s work had positive resonance in an era when study of the medieval past focused on the origin of American ideals, institutions, and customs, rather than its alterity, as was the case in the later twentieth century.⁶⁵

Thus, beyond seeing the work of Pirenne as a profound meditation upon how World War I caused the Belgian historian to renounce his formerly uncritical embrace of German scholarship,⁶⁶ medievalists should be mindful of the impact of Europe’s colonial relations with North Africa and the Middle East in the late nineteenth and early twentieth century on Mahomet et Charlemagne. Although the asymmetric ignorance of Pirenne’s approach should not surprise anyone, given the proclivity of many European scholars to produce theories that embraced large parts of humankind without learning much about the cultures affected,⁶⁷ Pirenne’s assumptions about post-Roman North Africa and the Levant followed in the tradition of a century of historical works that justified European colonization through reference to former Roman territories and their rightful restoration to the sphere of Europe.⁶⁸ Pirenne proclaimed that in the seventh century “a rupture occurred that will last to our days.”⁶⁹ There is thus no doubt that he viewed the modern age as deeply connected to its Roman and post-Roman past. In this line of thinking, history had come full circle. Just as Pirenne argued that Charlemagne could not have existed without Muhammad, he also understood the contemporary heirs of Charlemagne, namely modern Europeans, to have conquered the heirs of Muhammad and restored Rome once again. Given the realities of post-9/11 global politics, medievalists must be wary of relying uncritically upon Pirenne’s binary account of civilization and barbarity, East and West, and Christian and Muslim. Perhaps, as suggested recently by Richard Hodges, one solution is to start by enlarging the frame in which we view medieval European interactions. We may consider the modest appearance of Carolingian activities relative, for instance, to the rise of China, a topic that likewise has powerful resonance in modern global politics.⁷⁰

Archaeology and the Future of the Pirenne Thesis

Since Pirenne wrote Mahomet et Charlemagne in the mid-1930s, the area in which the landscape of post-Roman research has changed most profoundly is archaeology.⁷¹ Although archaeology, like history, is a discipline contingent upon the conditions in which its practitioners operate, it is the largest source of new data useful to improving our understanding of late antiquity and the early Middle Ages.⁷² Integrated with the historical data we possess for the post-Roman West and framed by anthropological theoretical approaches, it allows us to reconsider the transition between ancient Rome and the Middle Ages in a more complex fashion than was possible in the early twentieth century. While its practitioners are far from being in agreement about its methods and implications, post-Roman archaeology has the greatest potential to move medieval studies beyond the framework inspired by Pirenne’s vision of the processes that transpired in this epoch.⁷³

Archaeology’s value to historical questions, however, has not always been immediately apparent. In the late nineteenth century, and certainly as late as the interwar period, when Pirenne wrote, late Roman and early medieval archaeology was not yet professionalized and its practitioners were organized largely within the context of provincial learned societies.⁷⁴ Driven by regional and nationalist concerns, cemeterial excavations represented the predominant genre of post-Roman archaeological research in Western and Central Europe.⁷⁵ Typological studies of the contents of row graves fueled debate over the identification of the ethnic populations buried at these sites and had implications for both modern German territorial claims to regions like Alsace and Lorraine and Pirenne’s claims that little change was effected by the Germanic invasions.⁷⁶ However, cemeterial excavation reports—which were highly local in focus and mainly interested in the identity of deceased warriors—did not lend themselves easily to broader arguments, like that of Pirenne, dealing with production and trade. Few historians in Pirenne’s day, aside from Mikhail Rostovtzeff in The Social and Economic History of the Roman Empire (1926), availed themselves of the archaeological record in studying the ancient economy.⁷⁷

By contrast, classical archaeology, which was intrinsically linked to colonial concerns and the acquisition of imperial collections for metropolitan and colonial museums, focused for the most part on epigraphical remains, monuments, and statuary.⁷⁸ In the nineteenth century, its methodology was quite simple and consisted in many cases of clearing statues and edifices of the debris of subsequent centuries and studying, drawing, and collecting monuments and inscriptions with approaches borrowed from art history and architecture.⁷⁹ In addition, scholarly interest in the classical past was trained almost exclusively on remains dating earlier than the third century, while later material evidence was typically misdated, ignored, or destroyed in the process of excavation.⁸⁰ Driven by written sources, modern aesthetic sensibilities, and colonial concerns, classical archaeology in the Maghreb thus mainly addressed sites that predated the critical era that brought Roman decline.⁸¹ As such, it was not well disposed to discussing the vitality of the post-Roman Mediterranean. One of the exceptions to this trend, and one of the first European scholars to undertake studies of Byzantine archaeology, was Charles Diehl, who published, among other works, the pathbreaking Ravenne: Études d’archéologie byzantine (1886) and L’Afrique byzantine: Histoire de la domination byzantine en Afrique (533–709) (1896).⁸²

From the 1930s, however, archaeological attention shifted to larger-scale excavations of trading sites that might elucidate the existence of trade in the post-Roman period, especially outside the Mediterranean. These undertakings included, most famously, the contested exploration of Dorestad at the mouth of the Rhine and Haithabu near Schleswig.⁸³ These excavations, too, were colored by nationalist concerns but nonetheless produced important evidence of production and consumption.⁸⁴ In the 1940s and 1950s, excavations were organized at Birka, Southampton, Hamburg, Helgö, Kiev, and Novgorod, among other locations.⁸⁵ As observed by Richard Hodges and David Whitehouse, in the 1960s, the rise of New Archaeology (also called processual archaeology) brought greater openness toward scientific methodology and a willingness to engage theoretical approaches in the study of cultural change from data.⁸⁶ The development of new archaeological techniques, including field-walking surveys (a noninvasive method for systematically and collectively surveying plowed fields and recording surface materials) and aerial photography, which offered helpful perspectives on still-unexcavated field sites, generated large quantities of useful data. Postwar policies and funding that supported European integration likewise supported larger-scale and costly site excavations, which allowed for the confirmation or refutation of certain features of Pirenne’s confidence in the continuity of trade until the Arab conquest. Excavations in the city of Marseilles, for instance, suggested that there was thriving trade as late as the early eighth century.⁸⁷ The exploration of Siraf by Whitehouse likewise introduced into these discussions considerations of trade far beyond the Mediterranean, including Sassanid and Umayyad maritime activities in the Persian Gulf.⁸⁸

By the 1980s, advances in research had revolutionized archaeologists’ understanding of the late and post-Roman periods, and advocates of New Archaeology revealed that a number of Pirenne’s basic premises were deeply flawed. Rather than seeing Islam as a central catalyst for the decline of the Western Roman Empire, as had been argued by Pirenne, for instance, Hodges and Whitehouse proposed pushing the date of the fragmentation of the Mediterranean earlier. In their reassessment of Pirenne’s Mahomet et Charlemagne, they reintroduced the prospect of significant Western Roman decline in the sixth century, followed by the eclipse of Byzantine trade in the Mediterranean a century later. They reimagined the nature of trade from Pirenne’s narrowly monetary definition and built on the groundbreaking proposals of Bolin from the 1930s onward, since he effectively integrated Nordic economic activity into the dynamics of the Caliphate’s trade.⁸⁹ Using recent data derived from archaeological excavations, Hodges and Whitehouse documented in significant detail the flourishing of production and trade in northwestern Europe, then governed by a patchwork of rulers and monastic institutions. Rather than seeing Islam as severing ties between East and West, they understood the Arab world as contributing to its success during the Carolingian renaissance.⁹⁰ They argued that the end of this productive period was in the early ninth century: it occurred centuries after the Arab conquest and was caused by Viking attacks in the West and the collapse of the Abbasid regime.⁹¹

However, thirty years after the publication of the Hodges and Whitehouse assessment of Pirenne, the work of post-processual archaeologists has made it is easier to see the flaws not only in the work of Pirenne but also in the positivistic approaches advocated by New Archaeology.⁹² In applying broad-based theoretical models for the phenomena observed rather than assessing emporia individually, archaeologists made too many assumptions about royal power and long-distance trade. They effectively neglected the complexity of human agency as a factor in these historical developments.⁹³ Moreover, their analysis of historical texts lacked interpretive nuance and their projects largely reduced Pirenne’s sweeping vision of civilizational change to economics.

In addition to less positivistic theoretical approaches, recent decades have brought the application of scientific technologies that have nuanced at the same time that they have complicated understanding of post-Roman developments. Among these tools are high-precision radiocarbon dating (which measures the amount of ¹⁴C in organic matter and allows scientists to estimate the time elapsed since the death of the plant or animal) and more accurate dendrochronological sequencing of wooden artifacts (based upon tree-ring patterns) for this formative epoch. They contribute to scholars’ ability to establish more reliable relative and absolute chronologies for archaeological sites (i.e., chronologies that date deposits or events relative to one another versus those that fix these events to specific years based upon an independent measure like radiocarbon dating). They help, for instance, to rectify the undercounting of materials from periods like the fifth or eighth century, in which grave goods were not used as frequently or were not distinct enough to support typological analysis.⁹⁴ These techniques also allow archaeologists to test the accuracy of the dating of sites integral to understanding trade in the late- and post-Roman epoch, an improvement over traditional dependence on a combination of stratigraphy, artifactual styles, and relevant historical references, all of which have proved less reliable than once thought.

Likewise, the introduction of increasingly affordable scientific technologies, like isotopic analysis, has increased what can be learned from organic remains relative to travel and diet. Isotopic analysis of bones and teeth using oxygen and strontium signatures allows archaeologists to determine with some level of accuracy individuals’ place of birth; studying skeletal remains for carbon and nitrogen signatures reveals information about diet during their lifetime. When applied as a means of testing conclusions resulting from typological analysis of grave artifacts, the input of isotopic evidence, for instance, has challenged traditional archaeological hypotheses about the expression of identity in early medieval cemeteries.⁹⁵ Although still in its infancy, DNA testing, improved by next-generation sequencing (NGS) that works well on shorter, decayed strands of ancient DNA, has permitted scientists to derive more detailed information about the genetic traits, illnesses, and relations of individuals, families, and larger groups.⁹⁶ While imperfect in terms of the kind of evidence they can deliver, particularly in the case of the analysis of ancient DNA, which is very costly and relatively imprecise with respect to when specific migrations and mutations of genes may have occurred, these techniques enable scholars to refine some of the questions they ask about trade, urban life, and migration.⁹⁷ DNA tests sometimes confirm and sometimes challenge long-held assumptions about gender associations of artifacts in early medieval graves.⁹⁸ Other approaches useful to archaeologists who need to map ancient sites relevant to the discussion of post-Roman trade include non- or minimally invasive techniques, like geographic information system (GIS) mapping and global-positioning-system (GPS) plotting.⁹⁹

The mass of data that these techniques have produced has required, in turn, new means of processing this once-unimaginable wealth of information. The increasing power and accessibility of computers, which from the 1960s made data analysis possible on this scale, have enabled scholars to observe larger patterns of human behavior in the archaeological record than was previously possible.¹⁰⁰ Michel Bonifay’s recent discussion of the configuration of consumption in late antiquity, a study based on enormous quantities of ceramic shards, for instance, suggests that Mediterranean trade was restricted to a narrow strip of the coast of North Africa and did not penetrate deeply into the local and regional North African economy.¹⁰¹ However, the application of techniques derived from the laboratory and computer sciences to the finds of archaeological fieldwork has meant that archaeologists now typically work as members of large excavation teams, rather than in isolation, as was common in the past.¹⁰² These collaborations, which have enabled archaeological analysis to evolve dramatically, have nonetheless also brought new challenges as archaeologists bridge the disciplinary divides that separate them from pure scientists, who often enter into projects with different expectations, standards, and objectives from those of humanists. Indeed, scientific studies of early medieval data have been known to gain a life of their own in the rapid granting and publication cycles of the hard sciences, resulting in publications that lack the critical input of archaeologists and historians. Written quickly and devoid of humanistic nuance, these journal articles frequently contain simplistic and inaccurate narratives that grab headlines but are imbued with problematic claims reflective of the limits of their authors’ knowledge of historical realities.¹⁰³

Despite the fundamental critiques of many of the central tenets of the Pirenne Thesis from the perspective of recent archaeological research, its shadow over the field of early medieval studies has remained potent and continues to shape scholarly assessment of the late-antique economy. For the organizers of the European Science Foundation’s Transformation of the Roman World project (1993–97), a project that involved hundreds of scholars from across Europe (and a few from North America), the enduring attraction of the Pirenne Thesis was characterized by the way in which it bridged geographic divides as well as disciplinary ones. It was a useful framework for a European-wide project, even if, as Richard Hodges noted, Pirenne had little awareness of the material evidence available to him in his day.¹⁰⁴ Likewise, archaeologists have spent much energy, perhaps too much energy, using the recent influx of archaeological data to fill in the detail missing from so many of Pirenne’s nonspecific observations about the fate of individual classical cities that he observed largely from the perspective of the written sources.¹⁰⁵ Although archaeological data have helped document some of the basic fissures in Pirenne’s sweeping conclusions that were previously very difficult to document in detail,¹⁰⁶ historians and archaeologists alike have too seldom applied them to opening up truly alternative vistas of the post-Roman period.

In his weighty reassessment of the Pirenne Thesis published in 2001, Michael McCormick has focused on shifting Pirenne’s discussion from a discussion of international trade to evidence of communications. Taking advantage of the recent digitization of entire series of Latin editions, like those contained in the Patrologia Latina and the Monumenta Germaniae Historica, McCormick has taken his evidence from a broad array of written sources, numismatics, and archaeological evidence.¹⁰⁷ In seeking to expand Pirenne’s parameters beyond the trade of luxury goods, like gold, silk, papyrus, and spices, McCormick has included in his massive study, among other things, the movement of clerics, relics, ceramics, and slaves.¹⁰⁸ Despite his characterization of Islam as the stimulus of the Carolingian economy and his identification of the era of the eighth and ninth centuries as laying the groundwork for subsequent economic expansion (rather than as a period of decline), McCormick’s deference to the broad outline of Pirenne’s framework alters rather than replaces the grand narrative and thus has left in place some of the structural constraints imposed by Pirenne’s vision of the unity (and later disunity) of the Mediterranean.¹⁰⁹

McCormick’s work makes it clear that despite the increased sophistication of early medievalists’ research tools and the growth of data available for study of the post-Roman period, Pirenne’s now eighty-year-old metanarrative continues to orient economic studies of this transitional period. Mahomet et Charlemagne remains relatively intact, even when the concepts upon which it is based have changed dramatically in recent years from the positivistic approaches that were de rigueur even fifty years ago (and still continue in some circles to be accepted at face value). Buoyed by the belief that the beauty and simplicity of Pirenne’s vision are a virtue—what scientists refer to as parsimony and philosophers as Ockham’s Razor¹¹⁰—scholars of the post-Roman Mediterranean like McCormick have been content to let the larger framework stand intact as the nearest we may come to what Pirenne’s biographer Bryce Lyon characterized as the unsolvable enigma of the end of the ancient world.¹¹¹ And, indeed, Pirenne’s elegance has changed our thinking in ways that make it very hard to turn away from his model, however imperfect it may be.

Nonetheless, evidence has been building that medievalists would do well to avoid focusing on where the Pirenne Thesis is accurate or inaccurate and to move beyond its scaffolding in favor of new questions and parameters. With the benefit of more sophisticated archaeological approaches, greater sensitivity to scholars’ roles in constructing categories of analysis, and sustained interest in issues that Pirenne neglected as not being germane, the time seems long overdue to abandon scholarly attachment to the elegant vision of urban life and trade as a measure of premodern civilization. While no one approach will allow us to reconstruct accurately the many complex features of the post-Roman era, scholarly research must make more of the fact that significant regional and local variations characterized the transition from the classical to the medieval era.¹¹² Peregrine Horden and Nicholas Purcell, moreover, have advocated abandoning the primarily economic concerns that have become dominant since Hodges and Whitehouse published their archaeological assessment of the Pirenne Thesis. They have replaced it with what might be described as a civilizational approach inspired by Braudel, but one moving beyond the unifying effect that he and Rostovtzeff attributed to the shipping routes that crossed the Mediterranean.¹¹³ Their approach takes into account the great variability of climate, landscapes, agricultural techniques, religion, and a variety of human factors that existed around the Mediterranean while still acknowledging the importance of connectivity.¹¹⁴

Avoiding the kind of grand narrative espoused by Pirenne, the important contribution of Horden and Purcell offers fewer direct contentions than McCormick but also opens more doorways to further research.¹¹⁵ Their work suggests that even if scholars cannot completely avoid objectifying the past and making it their own,¹¹⁶ they can at least broaden their source base sufficiently to expose potential anachronisms in existing models. For those like myself interested in developments that occurred during the intervening Merovingian period, this approach might help undo the secondary status of the pre-Carolingian era that structural reliance on the Pirenne Thesis traditionally has fostered. Taken from the perspective of early medieval history and archaeology, these developments are already occurring in a number of interesting ways, the final subject to which I will now turn. This is very much a meditation rather than a complete survey of some of the enormous changes wrought in the last several decades by multidisciplinary research of the early medieval West.

Moving Forward: Big, Messy, and Multidisciplinary

In the last decade, the ways in which scholars think about the relationship between the Merovingian regions and the Mediterranean have grown significantly more complex as they have shifted to multidisciplinary approaches. Most prominently, Chris Wickham’s encyclopedic assessment of a vast range of historical and archaeological sources has altered the way in which scholars think about the period. Shaped by Marxist interpretive models rather than the Romanist outlook that characterized the writings of Pirenne, Wickham has updated and overhauled our understanding of the centuries from the late Roman Empire to the early Germanic kingdoms. Leaving behind the bold strokes of Pirenne, he has replaced them with a complex and messy pointillist synthesis. Instead of the Roman Empire, there are regions; instead of a single panorama, there are detailed and comparative case studies of urban centers, emporia, and rural estates.¹¹⁷ In the states that framed the lives of the landowners and peasants at the heart of Wickham’s study, there is continuity of preexisting regional differences in culture and economy, yet abundant elements of radical change throughout the former Roman world.¹¹⁸ The wealth and diversity of these parallel experiences—and the kind of microregions and connectivities likewise considered by Horden and Purcell¹¹⁹—are goldmines to plow through in graduate seminars with students. However, they are not easily boiled down to a digestible narrative to convey in undergraduate surveys in the course of addressing the great changes that characterized the late-Roman world. The anticlimactic conclusion that emerges from Wickham’s grand study, that “Charlemagne’s papal politics … would probably not have been significantly different had the eastern Roman Empire maintained its Mediterranean compass,”¹²⁰ would not have made headlines in even more humanities-friendly times.

Occupying the heart of Wickham’s study are comparatively modest ceramics, some fine and many coarse, some traveling great distances filled with wine, oil, and fish sauce, and others of local production and contents. These often-unattractive vessels have largely replaced the more glamorous (and visible) gold, silk, spices, and papyrus that were at the heart of Pirenne’s focus on Roman elites.¹²¹ Although the more orthodox Marxist critique by Jairus Banaji has convincingly questioned the optimistic vision of free-peasant production envisioned by Wickham and has proposed the widespread existence of slavery and landless laborers, the general picture Wickham proposes has held for the last decade.¹²²

In the future, however, Wickham’s conclusions about the end of Roman rule will hopefully be enriched by the synthetic approach taken by Frans Theuws, who has suggested innovative ways in which to account for the wide variety of grave goods found in contemporary early medieval cemeteries. Theuws has suggested that peasants made a more active contribution to the early medieval economy than acknowledged by Wickham.¹²³ Theuws has also effectively married the rich sources for religious and cultural phenomena in the early medieval period with the expanding wealth of material evidence for emporia, referred to in England as wics. These were the coastal or riverine trading outposts, some more permanent than others, that emerged in northwestern Europe between the sixth and ninth centuries. Understanding the ability of commodities and even coins to hold both practical and symbolic value,¹²⁴ depending upon the spheres in which they were exchanged, allows scholars to reconnect economic features of early medieval society with the imaginaries of their Christian inhabitants.¹²⁵ Alternative models, such as proposed in Peter Bang’s comparison of the Roman economy with that of the Mughal empire, or Anne Haour’s study of the similarities between wics and Sahelian trading centers, expand comparative possibilities for modeling the economics of premodern states.¹²⁶

Recent scholarship continues to push disciplinary boundaries by adding ecological concerns to examinations of the early medieval Mediterranean. The environmental projects of Paolo Squatriti, for instance, take on the weighty history of the chestnut, among other plants, and the impact of its cultivation on European diet and trade.¹²⁷ By reading between the lines of landholdings transmitted in charters, historical, and hagiographical descriptions of weather patterns and farming, and recipes in ancient cookbooks, he integrates these sources with dendrochronological, archaeobiological, and climatic research to challenge long-held notions of the early medieval landscape as wild due to the encroachment of what many have seen as dark and foreboding forests. He urges medievalists to reconsider the post-Roman stereotype that equated the spread of wilderness with a lack of civilization and turn instead to a more nuanced understanding of the active human participation that was required to cultivate the landscape.¹²⁸ With respect to the subject at hand, this work suggests, in turn, that medievalists cannot assume that disruptions in trade—specifically, in the subsidized transport of the annona—were the only factors that effected long-term changes in the European diet. That environmental changes were likewise important can be recognized from archaeological evidence of changing climatic conditions that either stressed or expanded the amount, location, and type of arable land in use at any one time.¹²⁹ At the very least, the integration of nontraditional sources into medieval studies allows scholars to create a system of checks and balances on evidence from narrative sources, which, although acknowledged as problematic in various ways, has too long dominated the discourse on the early medieval economy. This new evidence supports the development of a more nuanced, comprehensive discourse about civilization after the long diversion that was caused by the positivist liabilities of New Archaeology.

Moreover, other forms of analysis, which one might characterize as products of the scientific turn in archaeological research, have expanded the horizons of early medievalists to regions far beyond the Mediterranean which were nonetheless connected to medieval Europe. One such approach now in use is mass spectrometry, an analytical chemistry technique that allows scientists to test the chemical composition of small bone samples by bombarding them with electrons, ionizing them, and then measuring their mass-to-charge ratio. This technique has recently been applied to Viking Age combs found at the emporium of Ribe on the west coast of Denmark. The tests have allowed archaeologists to identify an increase in reindeer, rather than red deer, antlers as compositional components of Viking Age combs from the late eighth century. These finds suggest that increased demand for this raw material resulted in an active long-distance trade between central Jutland and the outer reaches of Scandinavia from the late eighth century.¹³⁰

In recent years, microspectrometry, a noninvasive technique, has also been used to analyze gemstones, among other substances, on the basis of the wavelengths of electromagnetic radiation they absorb or reflect. This tool has enabled archaeologists to identify the origins of raw materials used in luxury goods to demonstrate that the Mediterranean was not the sole conduit of long-distance trade for the Merovingian world; it opens up opportunities to identify items that traveled to and from distant points (identifications that were earlier achieved on the basis of stylistic analysis).¹³¹ The transmission of garnets—the basis for much of the red-colored inlay in high-status gold cloisonné brooches, buckles, and ornamented weaponry used by Merovingian elites, for instance—covered vast itineraries, reaching the West from as far afield as India and Sri Lanka until the late sixth century. Archaeologists have recently argued that during the seventh century, while this style of decoration still prevailed, these garnets were increasingly replaced with those arriving from locations in Bohemia and possibly other parts of Europe. Although the application of inset garnets to jewelry and other luxury items disappeared altogether within several decades of this transition as they were replaced by décor made of colored glass fragments,¹³² these studies suggest that the circulation of raw materials was considerably broader than the boundaries of the Roman world and depended upon Arab traders, who worked with considerable volumes of goods and covered enormous distances.¹³³ Direct trade with China, rather than through East Asian intermediaries, was also accomplished by the ninth century, if not before.¹³⁴

Nor are garnets an exception in this sense; in the eastern part of the Merovingian realm, ivory, presumably from India and Africa, and cowry shells from the Red Sea (Cypraea pantherina) and from the Eastern Mediterranean (Cypraea tigris) have been identified in early medieval Continental and Scandinavian burial contexts and as stray finds.¹³⁵ Glass, often still in raw form, likewise made long journeys.¹³⁶ Recent techniques, like fission-track dating of the minute quantities of uranium contained in glass, have made it possible to date specimens more accurately than on the basis of style alone.¹³⁷ In a recent study of excavations at over thirty cemeteries across Merovingian Gaul, including many sites in what is now Belgium, Constantin Pion’s detailed research has revealed the frequently distant places of origin of relatively modest glass beads found in these graves. The origins of these easily transportable objects, in a variety of colors, including yellow, black, orange, reddish orange, and green, have been traced not only to Mesopotamia but also to Sri Lanka and India. These objects were apparently popular among the populations that occupied northern Gaul primarily in the late fifth and early sixth century. From about 530 CE, however, more exotic glass beads were increasingly replaced by local production, leading to their disappearance in the first third of the seventh century. Like that of garnets, the cessation of the long-distance travel of glass beads was owed possibly to disruptions in the Byzantine world that made it difficult or impossible to continue conveying such desirable ingredients for popular, though not necessarily exclusive, decorative elements.¹³⁸

Beyond suggesting the scope of Merovingian trade, the distribution of beads at individual cemeteries (and, in a sense, the processing of “big data” required in the study of prolific quantities of small artifacts) also provides a possible means by which to modify existing relative and absolute chronologies of early medieval grave goods.¹³⁹ Beads, which typically occur in large numbers on early medieval grave sites, offer an additional framework by which to date objects that were formerly assessed on the basis of their stratigraphic proximity to identifiable coins or on the basis of stylistic typologies established over the last century and a half. Current discussions of beads follow on developments in the study of ceramic fine wares, like African Red Slip ware, the dating of which has been revolutionized in recent decades as it has become possible to establish larger and more reliable databases.¹⁴⁰ Research in the field and the laboratory has shown that imported vessels became models for local imitations or were replaced to some extent by nonceramic containers, like wooden barrels, developments that can only be understood with a more thorough analysis of the fabric, production, provenance, and use of these objects.¹⁴¹ This research has, among other things, revealed profound inaccuracies in coin-based chronologies.¹⁴²

Besides the alternative perspectives offered by new technologies and multidisciplinary approaches that bring together science with humanistic assessment, early medieval scholars have continued to modify and adapt anthropological models in their discussion of material remains. Some have penetrated beyond traditional analyses of the movement of goods from point A to point B and focused instead upon the detailed documentation of production sites, like the ninth-century Benedictine monastery of San Vincenzo al Voluturno in the province of Isernia, Italy, and the eighth- to ninth-century emporium Reric at Gross Strömkendorf on the Baltic Sea coast.¹⁴³ In addition to providing essential information about the kinds of products being created at the workshops related to these trading sites, archaeologists and some historians are thinking about the physical properties of the material—clay, iron, wool, and so on—from which artifacts were made and the restraints and affordances such properties imposed on those who worked with them.¹⁴⁴ Borrowing from the work of social theorist Igor Kopytoff nearly thirty years ago, they have nuanced their work to take into account the biographies of objects and the changed meanings and social relations they fostered from the time of their production to their final disposal or abandonment.¹⁴⁵

An example of such an approach, for instance, is found in the study of post-Roman recycling and the manner in which some former Roman regions coped with economic collapse from the fourth through seventh centuries. Robin Fleming’s ongoing assessment of the metal economy in post-Roman Britain from 350 to 650 CE, a topic directly related to Continental concerns in the post-Roman era, is suggestive of the adaptability and innovation of local smiths following the cessation of smelting of newly mined iron in the 370s. The recycling of metal elements from Roman structures became a common practice in Britain over the next several centuries, and awareness of this practice should make scholars wary of reading the post-Roman economy as functionally dependent, foremost, upon trade and upon the accumulation of wealth as traditionally assessed.¹⁴⁶ Changes in the manufacture of spears, as has been pointed out by Andrew Welton, demonstrate that recycling was not a desperate measure; our understanding of the method and materials used and symbolism embodied in the production of spears in post-Roman Britain must be more nuanced. Recycled iron, an important but uneven material, tested the skills of the smiths who used worked spears. The success or failure of their workmanship had a direct impact on the nature and quality of the military, social, and cultural interactions in which the spears played a role.¹⁴⁷ Influenced by the growing importance of materiality studies,¹⁴⁸ research on objects like combs has shown that even modest toiletry implements had layered meanings and a complex history.¹⁴⁹ These considerations should figure in and complicate understanding of the objects that have counted so large in assessments of local, regional, and long-distance trade.

Conclusions: Messiness as a Postcolonial Virtue

The milieu within which Pirenne operated and the war that changed him as a scholar remain in the living memory of no scholars alive today. The assumptions and privilege of the colonial Mediterranean that shaped historical knowledge of Pirenne’s day and affected his vision of the past, while certainly not erased, are at least now theorized and critiqued. The anachronistic residue in the Pirenne Thesis of historical, cultural, and religious assumptions that went largely unquestioned throughout his lifetime make more than just the details of his vision of the clash of civilizations problematic for the periodization of the ancient and medieval West. Homi Bhabha has observed, “An important feature of colonial discourse is its dependence on the concept of ‘fixity’ in the ideological construction of otherness.”¹⁵⁰ Pirenne’s emphasis on the fact that the Muslim Arab invasion of the eastern and southern ends of the Mediterranean forever cleaved it from the Christian West does not do justice to contemporary efforts to understand the post-Roman world and its broader connections. Like many other works of this era, Mahomet et Charlemagne is sown with nineteenth- and early twentieth-century colonial language. However, whereas most other works of the 1930s have languished, Pirenne’s has not, and his ideas continue to enjoy a place of honor at the start of the twenty-first century. Recognizing the powerful influence of this text in a contemporary context, medievalists should be critical of Pirenne’s anachronistic assumptions and their implications in studies of the post-Roman world.

Although it is impossible to see into the future with any clarity, there is no doubt that scholarly approaches shaped by the uncertain events of the early twenty-first century will be used in coming decades in unexpected ways. Historians and archaeologists will dissect this work in relation to the ideological expectations of their own day and will replace still-imperfect approaches with alternative interpretive strategies and analytic techniques. However discouraging it may be to recognize the cyclical nature of this scholarship, this cycle of reinterpretation should not prevent medievalists from exposing the way in which the contemporary perspective we have, in this specific moment in history, has colored our understanding of the medieval past. New findings and methodologies should inspire scholars to find more nuanced ways in which to conceive of the transformation of the Roman world. The time is overdue for us to embrace the uncertainty and messiness that comes from recognition of the complexity and variety of the late-antique and early medieval Mediterranean world in preference to our continuing to adhere to the simple beauty yet dangerous flaws of Pirenne’s politically charged grand narrative.