In the preface to the first edition of Casa-grande e senzala, Gilberto Freyre described his voyage from the west coast of the United States to New York, soon after finishing his book. Recalling Louisiana’s swamps and its fields of sugar cane, he wrote:
Every student of the patriarchal regime and the economy of slave-holding Brazil ought to become acquainted with the “deep South.” The same influences deriving from the technique of production and of labor—that is to say, the one-crop system and slavery—have combined here in this English-settled portion of North America, as in the Antilles and Jamaica, to produce social results similar to those that are to be observed in our country. At times, indeed, they are so similar that the only variants to be found are in the accessory features: the differences of language, race, and forms of religion.Footnote 1
Less than a century earlier, Henri Wallon, in his monumental Histoire de l’esclavage dans l’Antiquité, attempted to shed light upon the situation in the French colonies through an examination of Greco-Roman slavery:
Slavery in the ancient world! It may seem strange to seek it in the distant past, when slavery is still among us. In doing so, I do not seek to distract people from the colonial question; to the contrary, I would like to focus their attention upon it, and encourage them to find a resolution.Footnote 2
In fact, several of the seminal works on the history of slavery have been influenced, to varying degrees, by a comparatist perspective, as though an analogical digression were the only way to make the mechanisms of slave domination explicit—to make them “speak,” as it were.
Be that as it may, ought we to agree with Patrick Manning, one of the greatest specialists on the subject, that “the field of slavery studies has become a model for comparative study in social and economic history”?Footnote 3 This depends on what we mean by the term “comparatism,” which has come to denote approaches as different in their methods as they are varied, and even contradictory, in their goals. Essentially, we must avoid confusing the comparatism that aims to study the morphology of slave societies in all their temporal and geographical diversity with the comparatism that proclaims its alliance with global history, and attempts to analyze the dynamics that have united different systems of dependency and labor regimes across the world since the early modern era. These two approaches not only designate radically different fields of investigation; they also and above all present two heterogeneous conceptions of slavery.
Since the end of the 1990s, the global history of slavery has made a particular effort to highlight the epistemological naivety of a certain comparatist tradition that understands slavery from the perspective of its institutions, and not as a dynamic process that results in each case from specific historical conditions. However, it is worth identifying the limitations of this global approach when it claims to be the only method capable of defining slavery across history. Indeed, it is to the defense of a comparative history of slave societies, understood in all their morphological diversity, that this article seeks to contribute.
Although this type of morphological comparatism is widely practiced, it remains under-theorized, and texts that define what is at stake and clarify its methodology are few and far between.Footnote 4 In the form of edited volumes that gather contributions from various fields, it is often superficial and, even in the best cases, limited to simple juxtaposition, leaving readers to infer the broader synthesis on their own. Above all, and regardless of the quality of their individual articles, these multi-author volumes undertake the comparative exercise from an overarching perspective, drawing point-by-point contrasts between different slave systems while presuming their coherence. Because the scale of observation is often poorly defined, these works frequently end up confirming things that are already well known about the societies under discussion, as though the comparative exercise had no effect upon the nature of the questions that we pose about them.
Under what conditions might it be possible to write a new comparative history of slavery, one that can avoid both the well-worn conventions of naive comparatism and the inadequacies of the global approach? To begin with, we need an inventory of the theoretical problems that the study of slavery poses for contemporary historiography. Following that, I will attempt to show what a comparatist approach (once its methods have been freshly redefined) could bring to the study of one slave society in particular: Athens of the Classical period (fifth and sixth centuries BCE). My argument focuses on a specific organization of slave labor, whose essential traits were shared by numerous slave societies, in which a slave tied to the running of land, a workshop, or a commercial store made regular payments to his master. When approached from a comparatist perspective, the analysis of this configuration makes it possible to shed light on certain particularities of Athenian slavery that have long remained unclear.
The Comparative History of Slavery: An Inventory
At first glance, it seems that the comparatist approach has always driven a large proportion of publications on the subject of slavery. In the context of edited volumes, the desire to compare and contrast different configurations of slavery is manifest in many different forms, whether across the same geographical space,Footnote 5 different periods of time,Footnote 6 or even in regard to a specific theme.Footnote 7 Since its foundation in 1980, one journal has embodied this comparatist ambition more than any other: Slavery and Abolition. Presenting itself as “A Journal of Comparative Studies,” it rapidly attracted contributions from the most important specialists on the subject in nearly all the relevant geographical areas; less than two years later, its editorial committee included figures as diverse as Claude Meillassoux, David Brion Davis, Moses Finley, and Suzanne Miers.Footnote 8 Slavery and Abolition called for a comparatism without borders, whether thematic or geographical, which was just as interested in precolonial Africa as it was in the Greco-Roman world, and paid as much attention to forms of subservience in Asia as it did to Brazilian or Caribbean slavery. As the founding editorial put it:
Few subjects lend themselves more readily to comparative analysis than that of slavery, one of man’s oldest and most controversial institutions. … Slavery and Abolition will cover the shifting circumstances of slaves and slavery down through the ages—in the Far and Middle East, in Africa and in the New World, in ancient and modern times. It will provide a continuing retrospective of the phenomenon of human bondage—its psychological dimension; its historic, economic, and social aspects.Footnote 9
However, a quick glance reveals the particular interests and preoccupations of this comparatist field. The primacy of colonial slavery seems to have been a constant from the outset. Even when setting out to “compare without taboo,” the majority of collective works are most often focused on the different forms of slavery that emerged from the conquest of the Americas.Footnote 10 This should not be in the slightest bit surprising; this asymmetry is the product of a numerical imbalance between the different fields of study, which in turn cannot be disassociated from the social and political demands that surround the topic. Because of this dynamic, specialists on Greek antiquity or Ottoman slavery are generally invited to offer a counterpoint to the Americas, which appear to form the natural focal point for the phenomenon of slavery. Moreover, it is clear that the challenge issued by Slavery and Abolition at the moment of its founding is far from being met. Although the first issues placed an emphasis on the Muslim world, African slavery, and antiquity, the Atlantic very quickly became the focus of the majority of its articles. Thus, from 1981 to 2015, more than 80 percent of texts published in regular issues, plus twelve out of twenty special issues, were devoted to the colonial American world (the Caribbean, Brazil, and the United States).
Since the 1940s, there has been no shortage of single-author studies comparing the organizational forms of slavery, most often between two regions or two countries.Footnote 11 Nonetheless, a large part of comparatist research has been polarized by two nearly contemporary major works, both of which offered a synthetic perspective on slavery. Slavery and Social Death by Orlando Patterson and The Anthropology of Slavery by Meillassoux represent a major moment in slavery studies, and have long been regarded as touchstones for all consideration of the subject.Footnote 12 The aim of these works was synthetic, in that they each sought to provide a definition of slavery that could encompass its most disparate historical configurations, with colonial slavery ultimately occupying a rather marginal position. More broadly, both advocated an institutionalist approach centered on the rules and mechanisms organizing the reproduction of domination, as the relations of the slave mode of production can only be implemented “through organized action by members of society to create institutions; which establish and constantly renew the relations of production.”Footnote 13 The weaknesses of both approaches have been frequently pointed out. Some have drawn attention to the skewed empirical material that all too often led Meillassoux to extrapolate data beyond his usual discipline of African studies, while others have criticized the frequent imprecision of the information collected by Patterson. The definitions of slavery proposed by the two anthropologists have continually been amended since the 1980s.Footnote 14 Nonetheless, over time it is the institutionalist conception of slavery that has become the major focus of criticism.
Indeed, the synthetic comparatism of Patterson, Meillassoux, or Alain Testart seems today to be at once beyond the reach of a single researcher and out of place in a single-author project. If a comparatist flavor permeates a majority of single-author works, it is subtle and is primarily used to situate a specific subject within a larger context—or to adopt concepts and questions from other disciplines—rather than to systematically address a single question. Thus, without explicitly proposing a comparatist reading of Ottoman slavery, Ehud Toledano consistently borrows concepts that originate in other fields of the history of slavery, the most recent example being creolization.Footnote 15 In the same way, Peter Kolchin has advocated a “soft” comparatist approach centered on the study of a precise configuration, which is itself carefully situated within a much larger context. Contrasting this flexible comparatism to more systematic studies with generalist aims (to which he has also made significant contributionsFootnote 16), Kolchin sees it as the only means of including analysis of specific contexts, without which history would be nothing but a flattened sociology.Footnote 17 Yet if synthetic comparatism no longer seems appropriate, this is above all due to the fact that the comparative history of slavery has been utterly transformed by the global turn. Over the last twenty years, a stream of research has overhauled the analysis of slavery and the scales of observation at which it takes place, proclaiming the end of traditional comparatism and denouncing it as fundamentally ahistorical. In the wake of more general evolutions in social history, from which it cannot be disentangled, a new conception of slavery has emerged.
The History of Slavery and the Global Turn
With its history of forced migrations across multiple continents, its centrality to the production of globalized goods such as cotton and sugar, and its foundational role in the processes of diaspora and creolization, slavery is an indisputably crucial subject for global history. It is therefore hardly surprising that historians of slavery have played a central role in promoting global history, following Manning, a specialist on the Atlantic slave trade and pioneer of global history in the Americas.Footnote 18 It is also worth observing that one of the first uses of the term histoire globale in France occurred in a study of the African slave trade, written by Olivier Pétré-Grenouilleau.Footnote 19
In American slavery studies, this paradigm shift is bookended by two dates in particular. In April 2000, the American Historical Review published a “forum” dedicated to transversal approaches to slavery, or “Crossing Slavery’s Boundaries,” and invited Davis, one of the most important figures in the historiography of slavery, to write the opening essay. In his text, Davis contrasted traditional comparatist approaches as a whole to another, more promising methodology that considered “the interrelationships that constituted an Atlantic Slave System as well as the place of such racial slavery in the evolution of the Western and modern worlds.”Footnote 20 To reestablish a “big picture” of Western slavery, he argued, we must understand that it is inscribed in a history that has stretched over multiple centuries. This analysis must extend back to a rereading of the medieval legacies of European slavery and forward to a consideration of its effects upon the recent history of capitalism, as well as the modern discourse on freedom.Footnote 21 In January 2016, the first issue of the Journal of Global Slavery was published. In an editorial note, Damian Alan Pargas laid out the principles of this new perspective on slavery as a supreme “global and … globalizing phenomenon”: “Slavery can be considered a globalizing force because it by definition connected societies—indeed, connected continents—through warfare, power relationships, trade networks, and cultural exchange, not just in the age of the transatlantic slave trade but throughout world history.”Footnote 22 The journal’s founding confirmed both the existence of a specialized field analyzing the history of slavery within global history, and the importance of a “global paradigm” in studies of the history of slavery. Between 2000 and 2016, scholars such as Gwyn Campbell, Joseph Miller, and Michael Zeuske, to name just a few, produced numerous studies in this vein, contributing to the writing of a new history of slavery.Footnote 23
This new history has reshaped the global geography of slavery by bringing new focal points to light, centers of gravity which had slipped through the cracks of the broad boundaries traditionally drawn between civilizations. The Indian Ocean, for example, has emerged as a major space of trafficking—equivalent to between 15 and 30 percent of the Atlantic slave trade—with a key role in “putting the world to work” in the early stages of a world-economy in the nineteenth century.Footnote 24 A group of studies have used “connected history” to highlight how various forms of traditional dependency evolved after coming into contact with the European slave trade.Footnote 25 Between the Russian sphere, Central Asia, and the Ottoman Empire, early modern Eurasia has revealed itself to be a space across which the trade in human beings was accompanied by a transfer of different ways of organizing labor.Footnote 26 The history of Russian serfdom thus appears in a new light when compared with other forms of “forced labor” in the Eurasian region.Footnote 27 In a similar way, the various forms of trafficking and captivity present in the early modern Mediterranean are now seen as part of a broader geography and a long-term history, transcending the division between Christian Europe and the Muslim world.Footnote 28 From the perspective of enslaved populations, the global approach has led to an increased focus on the construction of identities that gave rise to specific cultures linking different continents, with the theme of diaspora occupying a considerable place.Footnote 29
Above all, the global approach has contributed to the establishment of a new paradigm in slavery studies, breaking with the analytical frameworks used by Patterson, Meillassoux, or Testart. This model is characterized by its refusal of a universalizing definition of slavery, criticized for being at once a ruinous abstraction and a problematic form of translation in regard to the multiple forms of dependency that are specific to all societies.Footnote 30 Instead, it is considered that historians ought to focus on the ways in which societies have continually conceptualized, in very diverse contexts, the extreme forms of domination and control which an overly superficial approach designates under the term “slavery.”Footnote 31 Rightly considering that this term cannot be used as a “ready-made” category by historians, this neonominalism implies taking it as one extreme potentiality within a continuum of situations of dependency.Footnote 32 In this vision, the constitutive polarity between free individuals and slaves specific to New World societies represents only one type of slavery-based relationship, and a rather unusual one at that; it ought therefore to be provincialized.Footnote 33
Even more than slavery in the strict sense, it is “the global system of forced labor” that lies at the core of this analysis.Footnote 34 Historians of the Middle Ages and the early modern period have long highlighted the particular historicity of the notion of “free labor.”Footnote 35 Following in their footsteps, numerous works of global history understand slavery as one modality among other forms of involuntary labor, alongside indentured servitude and serfdom. When viewed from this angle, slavery ceases to seem like a marginal phenomenon specific to colonies within a fundamentally Eurocentric world-economy—on the contrary, its rise is repositioned as central to capitalist modernity’s expansion.Footnote 36 Several conceptual propositions have emerged in this intellectual context. For instance, the notion of a “second slavery,” developed by Dale Tomich and Michael Zeuske to refer to a second serfdom, has come to designate the changes in the organization of slave labor that took place during the first decades of the nineteenth century, as well as its expansion and intensification in the context of the growing integration of colonial zones into the world-economy.Footnote 37
More broadly, the institutional dimension of slavery is central to the critique of traditional comparatism. Whereas Patterson and Meillassoux proposed a sociology and an anthropology of slavery, it is implied that a specifically historical interpretation of the phenomenon would involve thinking through, for each individual context, the reasons a society used slaves.Footnote 38 Rather than slave systems, researchers should consider this recourse to slavery, understood as a dynamic process that always relates to historically situated strategies. Miller thus prefers to focus on what he calls “slaving strategies”: if we accept that the legal institutionalization of slave domination would historically have been the exception rather than the rule,Footnote 39 the taxonomical distinctions that have long guided analysis of various slavery-based organizations are largely unhelpful.
The notion of slavery as an institution has thus been replaced with a notion of slavery as an “involuntary relationship of mutual dependence between two quite unequal partners.”Footnote 40 And because “slavery was not merely what the masters wanted it to be, but a historically changing relationship that was also shaped, if asymmetrically, by the slaves’ wishes, strategies and identities,”Footnote 41 the challenge is to shed light on forms of servile agency, which an institutionalist reading can only conceal. In this sense, at least, the global history of slavery, particularly in its American forms, is not unlike subaltern studies.Footnote 42
The Limits of the Global Approach: An Ancient Historian’s Perspective
This global approach to slavery has undeniably had much success. In the best cases, it avoids the criticisms that have often been made of global history, accused of evoking too quickly “overarching identities and entities with capital letters (‘China,’ ‘State,’ ‘Capitalism,’ ‘Europe’).”Footnote 43 By making detours through Sokoto or Central Asia, it deconstructs linear narratives about the modernization of the world and subversively but insistently traces unsuspected links between capitalist modernity and slavery. Furthermore, it deconstructs objects of study that, upon examination, are revealed as historiographic chimeras: it is no small thing that this global history, among other achievements, has shown that “Islamic slavery” is a meaningless notion.Footnote 44 When it manages to operate on multiple scales of observation, even building on the achievements of microhistory, it contributes to a profound renewal of our understanding of colonial societies.Footnote 45
However, it is incumbent upon us to ask what this global approach may have missed regarding slavery, given its claims to provide the only reading that is properly historical. One obvious point, which has been made by many, is that global history embraces the diversity of the world using only one idiom and a preferred terrain: English, and modernity in the broadest sense. This is less a question of individual researchers’ skill sets than of a “global” bibliographic culture, whose echo chamber—now readily accessible on multiple platforms—inexorably consigns whole swathes of knowledge to oblivion, their sole sin being that they were written in a minority language.Footnote 46 In this way, the French-language tradition of African studies, though it produces pioneering research on the subject, has been gradually relegated to the bibliographical background in favor of exclusively English-language references; the same can be said of the colossal Portuguese canon, which radically transformed the analytical frameworks used by Freyre, to give just one example. Another subject that begs discussion is the oft-repeated reference to the “agency” of slaves, which is ethical as much as it is scientific,Footnote 47 and can sometimes lead to serious historiographical misinterpretations.Footnote 48 For although it is wrong to make slaves the passive instruments of their masters’ will, it must be recalled that slavery is by its very nature non-negotiable, and that the hegemony of slave domination involves, at least in legal terms, the negation of the slave as a person.
Yet this is not the most essential point. Above all, we must observe that this global history—centered on a modernity that extends from the earliest colonial contact to the post-abolition period—has some blind spots. First of all, certain geographical zones—such as early modern China or the Ottoman world, both domains in which the study of slavery has witnessed some important developmentsFootnote 49—pose difficulties when it comes to integrating them into the “big picture” called for by Davis. Another omission, just as striking, concerns the place of the ancient world in general discussions of slavery. Although it lay at the core of Davis’s early research, and even though Patterson himself used it as a matrix,Footnote 50 Miller’s manifesto on the global approach to the history of slavery devotes only a few pages to the subject, and these are very general and often problematic.Footnote 51 It would seem that Greco-Roman antiquity has nothing to contribute to the global history of slavery, either in terms of method or in terms of content.
Such an omission is not surprising, insofar as it is difficult to apply the theoretical framework constructed by global history to Greco-Roman slavery. This is not simply because slavery was integrated into all levels of production in the economy of the ancient Mediterranean, rather than confined to its peripheries. Above all, the “inclusivist” approach to slavery, conceived within a spectrum of forms of dependency, is ill-suited to describing societies that constructed freedom and servitude as diametrically opposed. In the world of the Classical city, individual status played out against a backdrop that was organized around a fundamental logic of exclusion and inclusion: slaves and citizens did not make up a “society.”Footnote 52 In this sense, slave status was not a position hierarchically inferior to the lowest censitary category of Athenian citizen. It derived from a radical exclusion, and was the product of an unequivocal domination.
Fundamentally hegemonic and uncontested in Greco-Roman culture for nearly a millennium, domination over slaves was well and truly an institution in the broadest sense of the term: without it, the fundamental elements of social reproduction would have been impossible. However, it also depended upon institutions in the narrower sense—that is to say, upon a set of coercive norms which ensured its survival. It is precisely in the recognition of this institutional aspect of slavery that the comparatist approach has something to offer. It is perhaps less important to define the status of a slave in every time and every place, which can only lead to unsatisfying formulations, than it is to construct pertinent analytical tools that enable us to identify different societies about which we can pose the same questions.Footnote 53
Slavery as an Institution: A Return to the Concept of “Slave Society”
More than any other scholar, the work of Moses Finley has embodied the institutional approach to slavery. In his article “Slavery,” published in the International Encyclopedia of Social Sciences in 1968, he was the first to make the enduring distinction between “societies with slaves” and “slave societies.” This distinction had a dual purpose: on the one hand, it aimed to describe the specific characteristics of a certain number of authentic slave societies, of which there were only five in all of history (Greece and Rome, the West Indies, Brazil, and North America); on the other hand, it claimed to offer an alternative to the Marxist theory of “stages.” It was thus to be understood in terms of a mode of production: if slavery is an “institutionalized system of large-scale employment” and “if the economic and political elite depended primarily on slave labor for basic production, then one may speak of a slave society.”Footnote 54
Today, Finley’s distinction seems problematic, not least because it is Eurocentric. It is clear that the societies of Classical antiquity and the New World are not alone in their unhappy claim to being slave societies.Footnote 55 Moreover, by focusing on labor, this distinction makes it difficult to explore the different reasons that societies have resorted to slaves at certain historical moments. However, it is one thing to criticize how the distinction is applied, and quite another to refute the heuristic value of all typological distinctions between societies that have made use of slaves.Footnote 56 Provided that its parameters are carefully redefined, the concept retains its utility, even offering a springboard for the comparatist exercise.Footnote 57
Contrary to popular opinion, Finley did not invent the notion of “slave society.” In reality, its history is at best a tortuous one, as its earliest instances were in the work of anti-abolitionists. It was first used in the mid-nineteenth century by George Fitzhugh, to oppose the “slave society” of the American South to the “free society” of the Northern States.Footnote 58 In Fitzhugh’s racist sociology, this term conferred a moral authority upon the Deep South, the so-called liberal society of the North being just another name for subjugation to the power of capital.Footnote 59 Its anti-abolitionist origins are the likely reason that the term fell out of use and only reappeared a century later, scientifically neutralized in the work of Frank Tannenbaum and Elsa Goveia, from whom Finley borrowed it, though not without adjusting its meaning.Footnote 60 For Tannenbaum, the term did not serve to distinguish between two forms of society, but instead shed light upon the centrality of slavery and its totalizing nature. In a slave society, he argued, the institution of slavery had its own logic, and overdetermined all social relations:
Wherever we had slavery, we had a slave society, not merely for the blacks, but for the whites, not merely for the law, but for the family, not merely for the labor system, but for the culture—the total culture. Nothing escaped, nothing was beyond or above or outside the slave institution; the institution was the society in all of its manifestations.Footnote 61
Goveia, for her part, distinguished between “slave systems” and “slave societies”: while the former designates the organization of slave labor and its economic laws, the latter qualifies the entirety of social relations which are conditioned, even indirectly, by slavery. We ought to focus, not on slavery’s place in the organization of labor, nor on the more or less considerable demographic presence of slaves in a society as a whole, but rather on the decisive nature of slavery as an institution across the different levels of the social experience.
When employed in this way, the notion of “slave society” retains all its capacity to discriminate between different kinds of societies that have made use of slaves, even beyond the Atlantic contexts in which Goveia and Tannenbaum were working. It opens a field of analysis to the comparatist approach that is at once sufficiently broad (it goes beyond the limited frame of the “Western world” or colonial societies) and precise (not all societies composed of slaves were slave societies). Above all, this notion allows us to analyze the institutional forms, in the largest sense of the word, through which slave domination perpetuates itself. A society can be designated as a slave society not because it has a significant number of slaves, nor even because it is heavily dependent upon slave labor for its production base, but because the use of slaves is indispensable to its ability to reproduce itself in various dimensions (productive, demographic, military, symbolic), making it the object of a conscious choice. This is far from obvious.Footnote 62 In the Ottoman Empire of the second half of the nineteenth century, slaves represented only 5 percent of the population. The determining characteristic of the institution of slavery, at least in the heart of the imperial space, did not depend upon either its demographic or its productive dimension. Nevertheless, because of the decisive nature of the functions with which slaves were entrusted in the administrative and military spheres, and the crucial role played by the institution of the harem, it is clear that the most essential elements of this society could not have reproduced themselves without making use of slaves. If the Warring States period in China (481–221 BCE) saw the emergence of what could be qualified as a slave society, this was not due to its slavery-based “mode of production.” It was rather that the introduction of chattel slavery overturned all forms of dependency, creating an extreme of radical exteriority whose very existence reshaped all social relationships over the long term. The new possibilities offered by slavery would moreover be at the heart of the unification work carried out under the Qin dynasty.Footnote 63
Toward a Medium-Range Comparatism
What comparatism should we, then, construct? It is striking to observe the extent to which the global approach is typically presented by its partisans as the natural extension, if not the necessary outcome, of the comparatist approach, typically using a reference to Marc Bloch as a smokescreen.Footnote 64 In 1928, Bloch distinguished between two types of approach. The first was founded on the observation of structural similarities between societies separated in space and time. Bloch essentially attributed three virtues to this method: “the possibility of filling in certain gaps in documentation by means of hypotheses based upon analogy; the opening up of new avenues of research suggested by the comparative method; above all, the explanation of a great many survivals that have up to now been incomprehensible.” Drawing parallels between this approach and Antoine Meillet’s general linguistics, he described it as a “long-range comparative method.” Though Bloch admitted that it could restore “with a kind of mental shock this sense of the difference, the exotic element, which is the indispensable condition for a balanced understanding of the past,” he nevertheless preferred another comparatist approach, centered on the study of “societies that are at once neighbouring and contemporary, exercising a constant mutual influence, exposed throughout their development to the action of the same broad causes just because they are close and contemporaneous, and owing their existence in part at least to a common origin.”Footnote 65
The comparatism that global history can claim as its own clearly corresponds to the second or “genealogical” definition, and not to the first or “morphological” one.Footnote 66 Indeed, Bloch’s reservations, critiquing the “search for similarities” and the “false similarities, which are often merely homonymous,”Footnote 67 regarding the comparison between English villainage and French serfdom in the late Middle Ages, are similar to those of Miller or Kolchin regarding the works of Patterson or Testart, whom they suspected of using the term “slave” to designate fundamentally heterogeneous kinds of dependency. But although the program set out by Bloch has been taken up by relatively few historians, other disciplines, chief among them anthropology, have partially achieved this project for a morphological comparatism.Footnote 68 Since the 1950s, under the influence of structuralism or at least within its orbit, the most innovative comparatist research has come from historical anthropology. Over the last twenty years, this comparatism has undergone a theoretical renewal, for which Marcel Detienne’s Comparing the Incomparable sought to provide a manifesto, even if it did not offer a serious discussion of method. Presenting itself “as a field of exercise and experimentation of the entire gamut of cultural representations among the most distant societies of the past as well as those closest to hand, and groups of living human beings observed on our planet both yesterday and today,”Footnote 69 it is clearly an example of the morphological comparatism evoked by Bloch. The weak points of this manifesto have been noted, in particular the fact that several of its propositions are unsuitable for social history.Footnote 70 Nevertheless, in a more or less critical dialogue with Detienne’s work, several proposals have been made from within ancient historical anthropology, which could potentially create a specific comparatism for social history that would be at once experimental, constructive, and reflexive.Footnote 71 These proposals join the work carried out around “thinking by cases,” where the application of a comparatist method involves subjecting a “case” to methodological reduction in order to isolate the unique context to which it belongs.Footnote 72
Resistant to grand taxonomies and “not concerned with exhaustiveness,” this comparatism above all seeks to identify “unique and essential elements organized into a variety of arrangements.”Footnote 73 Its goal is not to establish any kind of system into which each society would be sorted, but to implement a principle of specification in order to situate and identify an issue in the context of each of the societies being studied. Comparisons can thus be constructed not based on a presumed difference between slave systems (as in so many edited volumes that contrast systems, and even civilizations, while assuming their essential coherence), but by examining specific situations that have been carefully desingularized.Footnote 74 There are many specific configurations that lend themselves to this comparative exercise: a slave flees and finds temporary refuge in a place (a Greek sanctuary, Buddhist monastery, royal residence, or a “free village” during the French colonization of West Africa) before his or her fate can be decided; a master rents out some slaves, or sometimes they rent themselves, to another master, at Saint-Anne in nineteenth-century Guadeloupe, first-century Ostia, or sixteenth-century Istanbul; a slave testifies before a judiciary body, whether a county court in Alabama in 1820, the royal court of the Emperor of China in the second century BCE, or the Athenian Heliaia in 340 BCE.
Such a comparatism would be constructivist and reflexive, developing its own categories without turning them into objects of analysis that then define the universal character of a category in themselves. The typification that makes comparison possible can only ever be provisional and instrumental, in the sense that the categories of analysis are continuously reinterrogated and modified according to the results of the comparative exercise. Their function is not to construct ideal types, but to develop instruments that can explain a localized configuration.Footnote 75 It is thus crucial to establish a comparatist protocol. How can we avoid the “forced analogies” that Bloch warned against? In fact, this is a question of scales of observation. The failings of comparative histories are often due to the choice of a macroscopic scale: all too frequently, by attempting to contrast societies in their totality, comparative histories of slavery tend to lose themselves in generalities, which ultimately means that we only observe what we already know about each of the societies in question.Footnote 76 In contrast, the choice of a micro scale, capable of isolating objects specific to the institution of slavery in order to compare them, enables us to shed new light on, or rather to deploy in a new way, the unique aspects of each slave society. This renewed morphological comparatism thus seeks to inscribe itself in the extension of the lessons learned from microhistory.Footnote 77
It would be absurd to claim to present an entire comparatist project in the context of a simple article, and the project that I am calling for needs to be undertaken collectively. I would however like to show how my own field of expertise—slavery in Classical and Hellenistic Greece—could be enriched by such an approach. In particular, I will focus on one precise configuration of slave labor: a slave attached to the running of an area of land, a workshop, or a commercial store and paying a regular sum of money to his master.Footnote 78 Though this configuration seems at first glance to be an exceptional example of the powers of agency and negotiation that slaves could, on occasion, exercise, it was a specific institution inherent to numerous slave societies. Studying this configuration deepens our understanding of Athenian society by revealing the complexity of the ways in which slave labor was organized; it enables us to shift the terms of the historiographic debate concerning the contractual capacities of slaves; and finally, on the basis of a specific case study, it invites us to revisit wide-ranging questions that relate to the very foundations of Athenian slavery.
The “Settled” Slave: Comparatist Variations
Let me begin by considering three different kinds of organization. In the nineteenth-century Sokoto Caliphate, which stretched from the Niger River to Lake Chad (north of present-day Nigeria), murgu was a regularly scheduled payment made by a slave charged with the running of a property owned by his master, who entrusted the slave to maintain it.Footnote 79 Properly speaking, murgu was outside the purview of the contractual emancipation defined as muhataba in Islamic law.Footnote 80 This institution took many different forms in countries under Islamic rule. In the Ottoman Empire of the classical period, muhataba referred to a contract concluded between a slave and his owner, in which the slave promised to pay a regular sum linked to the running of a workshop or a commercial store in exchange for emancipation at a future date. The length of the contract was frequently set at four or five years, but could be as long as sixteen years.Footnote 81 Generalized use of this sort of arrangement could lead to serious ambiguity between the position of slave and that of salaried worker: instead of employing a free worker whom he compensated for a defined period of labor, the master bought a slave whom he freed at a future date, having established a muhataba contract connected to the administration of one of his properties. On the other side of the Atlantic, in the major cities of nineteenth-century Brazil, escravos de ganho paid a regular sum to their masters on a daily or a weekly basis. Able to possess a peculium independent of their masters, they were truly autonomous when it came to conducting their affairs, and this autonomy was recognized by the law. The contracts they drafted in their own names were legally valid, and their activity was recorded by municipal authorities, who alone could give or deny them permission to work.Footnote 82 In this context, the coartação was a widely used oral or written contract that allowed a slave to ultimately buy his freedom in exchange for the regular payment of a fixed sum of money.Footnote 83
Though any genealogical relationship between these cases remains relatively unlikely,Footnote 84 all three involved an agreement between a slave and a master, which was legally valid and could form the basis for a court appeal. The Islamic tribunals of the Sokoto Caliphate recognized the full legal validity of the murgu contract, in spite of this institution’s probable origins in Fulani common law.Footnote 85 If a coartação was not respected, it could lead to an appeal before the governor and thus to a trial, just as a master who concluded a muhataba contract with a slave was legally obliged to free him. However, the agreements made in each of these configurations had different end results. While the slave who regularly paid his master murgu remained a slave in the full sense,Footnote 86 the coartação or muhataba-style contract ultimately entailed his liberation. But even though these last two contracts also contained specific clauses stipulating that the slave could not be rented or sold during the period preceding his definitive emancipation, the contractual nature of the relationship in no way altered the slave status of the individuals involved. In this sense at least, they were legally valid contractual relationships that bound a master and a slave.Footnote 87
Athenian sources of the Classical period reveal numerous cases of slaves who paid a regular income (apophora) to their masters.Footnote 88 However, this payment in no way established a particular category of slave defined in law: it was entirely a means of organizing slave labor.Footnote 89 It is moreover possible to identify a specific kind of organization in which a slave was tied to the running of a store or workshop in which he had been settled (kathēmenos), without it necessarily being linked to the payment of an apophora. It seems that the placing of a slave was a distinct action on the master’s part.Footnote 90 The plaintiff of Demosthenes’s Against Pantaenetus thus accuses his adversary in the following terms:
Read what follows. “And after I had become a debtor to the state, he stationed his slave Antigenes in my mining property at Thrasyllus, in full control of my works, although I forbade him … (καταστήσας Ἀντιγένην τὸν ἑαυτοῦ οῦἰκέτην εἰς τὸ ἐργαστήριον τὸ ἐμὸν τὸ ἐπὶ Θρασύλλῳ κύριον τῶν ἐμῶν, ἀπαγορεύοντος ἐμοῦ).”
Stop reading. In all this he will again be convicted of falsehood by the facts themselves; for he has written in his complaint that I stationed the slave and that he forbade me (γέγραφεν γὰρ καταστῆσαι μὲν ἐμέ, ἀπαγορεύειν δ᾽αὑτόν). But this was impossible, as I was not in the country. I did not station anyone, seeing that I was in Pontus, nor did he forbid a man who was not in Athens.Footnote 91
The passage suggests the existence of a document that indicated the place (in this case, a property in the deme of Thrasyllus) and the name of the slave concerned. Yet there is nothing to imply that such a document entailed the recognition of a legal responsibility, even a very limited one, on the part of the slave. To the extent that such a text existed, its primary function was undoubtedly to establish for a third party a chain of liability between the “settled” slave and his master.
Two cases of slaves paying a regular fee to their master (kathēmenoi) have been extensively discussed in the literature, and deserve to be reexamined from a comparatist perspective. In 364/363 BCE, the inheritors of Pasion’s bank seem to have concluded a rental agreement (misthōsis) with four of their slaves:
To prove that I speak the truth in this, and that Apollodorus subsequently leased the bank to Xeno, Euphraeus, Euphro, and Callistratus, and that he delivered no private capital to them, but that they leased only the deposits and the right to the profits accruing from them, take, please, the deposition which proves these matters, and proves also that he chose the shield-factory. … Evidence has been submitted to you, men of Athens, that [the heirs] granted a lease to these men, and gave over to them no private banking-capital; and that they gave them their freedom, as if they had received great benefits from them.Footnote 92
These four slaves also had to pay their masters an annual rent of one talent.Footnote 93 The situation is clear: the two inheritors had agreed to lease money and goods deposited in the bank to several of their slaves, who were subsequently freed.Footnote 94 Xenophon also mentions the existence of a Thracian slave named Sosias, who during the Peloponnesian War (431–404 BCE) was entrusted by Nicias with the management of a thousand slaves who worked in the mines of Laurion. These slaves became the objects of a lease (misthōsis) between the two men, which was to earn Nicias one obol per slave per day.Footnote 95
How should we interpret these two situations, which ancient sources insistently present as lease agreements between slaves and their masters? Historians have for the most part been split between two camps. Some, who consider misthōsis a true contract between a master and a slave, conclude that the slave’s contractual capacity was fully recognized by the law. The lease agreement would thus have made the slave responsible for activities undertaken in the course of running the bank or the mine.Footnote 96 In contrast, others believe that the very principle of a lease precludes the possibility that slaves were involved: these men must have been freed before being entrusted with such a responsibility.Footnote 97 This last interpretation is, however, at odds with a close examination of the texts: it is clear that the four slaves of Pasion’s bank were freed after the misthōsis was contracted, and the slave status of Sosias the Thracian is never in question.Footnote 98 It must be conceded that an agreement or convention had been passed between master and slave. But should this be considered evidence of a contractual capacity that was legally recognized? There is reason to doubt this.
The comparatist approach reveals the flaws in this debate: on the one hand, the assumption that there is an ontological opposition between free labor and forced labor, with no intermediary forms of organizationFootnote 99; on the other, the idea of a fundamental contradiction between the order of status and that of the contract, stemming from the evolutionist belief that a contract represents “the end point and culmination of a historical progress, in which people were plucked from their subjection to status, and placed on the path of freedom.”Footnote 100 Each of the cases mentioned above demonstrates, without the slightest ambiguity, that slaves and their masters could enter into an agreement, though its legal force would naturally and consistently vary from one society to another.
The comparatist approach also offers the “possibility of filling in certain gaps in documentation by means of hypotheses based upon analogy.”Footnote 101 In light of the existence of muhataba- or coartação-type contracts, we could even wonder if the misthōsis between Pasion and his slaves contained the promise of an eventual emancipation—the existence of which, while not attested for Classical Athens, cannot in principle be ruled out.Footnote 102 Nonetheless, in relation to these two types of contract the Athenian case presents a striking peculiarity: there is nothing to suggest that a slave could have evoked the existence of such an agreement before the city’s courts.Footnote 103 Its significance seems to have been limited to the relationship, by nature patrimonial, that linked a master and his slave.
Master and Slave: Figures of Liability
A comparative perspective, initially centered on a specific and defined configuration, thus encourages us to ask a much broader question that goes to the heart of the relationship between master and slave. The degree and limitations of the former’s liability in relation to the latter’s activities was a question posed in all slave societies. It was at the center of arrangements like murgu, muhataba, and coartação, which, independently of any ultimate emancipation, allowed the master to delegate some of his own liability to his slaves. Looking at these arrangements from a comparatist perspective can shed light on the normative foundations of Athenian slavery as an institution.
Let us consider, as a starting point, the Insulindian world of the early modern era. Though the laws of Malacca, instituted in the mid-fifteenth century, make no mention of cases in which slaves acted as agents for their masters, the Undang-undang Kerajaan, the late sixteenth-century legal code of the Sultanate of Pahang, presents a very unusual configuration.Footnote 104 In article 11, it states that a master could not be held liable for the commercial debts contracted by his slave. Furthermore, it distinguishes between two categories of slave (hamba), from whom a lender could not claim the same rights. While a creditor was authorized to issue a loan to a slave who owned property, no loans exceeding the sum of one paha to a “destitute” slave could lead to any sort of compensation, including from the master of the slave.Footnote 105 Thus, the lender could not hold the master liable for obligations resulting from a contract with a slave. But here the absence of liability on the master’s part was associated with the recognition of the slave’s right to own property: it was against the slave’s own possessions that free lenders could enforce their claims, with the law imposing a clear limit in the case of “destitute” slaves.Footnote 106 There were no legal procedures regulating the respective liability of the master and the slave toward a third party, and any debt contracted with a slave led to an acknowledgement of his or her existence as a legal subject.
The Malaysian situation was a case of hybridization between Malay common law and the Shafi‘i school of Islamic law.Footnote 107 In contrast, the Mālikī tradition, as it was applied—with extensive local variants—in a significant number of Sahel-Saharan societies during the eighteenth and nineteenth centuries, stipulated that a master was entirely liable for all debts incurred by his slave, who in reality could not act except in his master’s name.Footnote 108 This unlimited liability meant that the slave was entirely an extension of the master. Nevertheless, beyond this general principle, societies based on Muslim law gradually developed legal means through which slaves’ liability in commercial actions could be recognized, while in parallel limiting that of their masters. Alongside the system of muhataba, in the Ottoman Empire of the classical period the ma’dhun slave could act as an agent by virtue of an authorization (idhn) granted by his master. This meant that he could conclude contracts on his master’s behalf but also for himself, as if he had regained legal capacities that had been only temporarily suspended by slavery.Footnote 109
This legal arrangement, which was a major part of economic life in the Ottoman Empire, naturally calls to mind the Roman system of preposition, through which a master could transfer his own liability to a slave, while placing limits on the extent of that slave’s actions, by drafting a lex praepositionis and presenting it to third parties.Footnote 110 The Pretorian edict titled Quod cum magistro navis institore eove qui in aliena potestate est negotium gestum erit—which can only be reconstructed using the later commentaries of Ulpian, Paul, and Gaius—aimed precisely to regulate disputes arising from transactions made by a dependent on behalf of his master, by defining the liability of each individual in regard to a third party. Between the third and the first century BCE, Roman law thus instituted six legal measures regulating the rights of a third party who had undertaken a transaction with a principal through the intermediary of an agent, whether or not this agent was a slave. Though the order in which these different procedures were introduced is a subject of particular debate,Footnote 111 three of them specifically concerned cases in which a slave served as the principal’s commercial agent. Among them was the actio de peculio, whereby the principal’s liability for the acts of his dependent was limited to the amount of the latter’s peculium, explicitly instituting a principle of limited liability on the part of the master.Footnote 112
It is not difficult to gauge the differences between the situation of the Athenian “settled” slave, and his Roman or classical-period Ottoman counterpart. It is unlikely that, in Classical Athens, the slave’s peculium ever had a legal definition—in the sense that Ulpian characterized it, following Tubero, as “the property which the slave, with his master’s permission, keeps in a separate account of his own, less anything owed to the master.”Footnote 113 In any case, it is certain that there was no procedure in Athenian law to limit the liability of a master in regard to his slave. The decisive proof is furnished by the case of Athenogenes’s slave, Midas, known from Hyperides’s Against Athenogenes. Athenogenes had entrusted the management of a perfumery to Midas, who contracted numerous debts in the course of this commercial activity. The plaintiff Epicrates, who had fallen in love with Midas’s son, bought both the son and Midas, but not before signing a contract with Athenogenes stipulating that Midas had incurred debts. These debts had, moreover, been contracted in the slave’s name, and not in his master’s.Footnote 114 Upon reading the whole plea, however, one can be in no doubt that Midas’s debts could only be imputed to the two owners, and not to the slave himself: in this sense, there was no limit to the master’s liability. Yet the sale of the slave did not nullify the obligations of the previous master to the lenders, thus necessitating a contract. By entering into a contract with the buyer Epicrates at the moment of Midas’s sale, the seller Athenogenes released himself from the debts contracted by his slave,Footnote 115 to the extent that the plaintiff’s entire argument rested on the question of the contract’s validity, which seems to have been rather unclear.Footnote 116 In this sense, the master’s liability for the actions of his slave was quite unlimited, as indicated by a law attributed to Solon and cited by the plaintiff: “Whatever losses and expenses slaves occasion shall be discharged by the master for whom the slaves are working.”Footnote 117
A specific procedure certainly allowed a citizen to pursue a slave in court, but this was only a preliminary phase in a legal action that was ultimately aimed at the master, who was entirely liable for his slave. The plaintiff of Demosthenes’s Against Pantaenetus could thus declare:
He ought not to be stating these charges now, nor to have inserted them in the challenge in which he demanded the slave for torture, but to have instituted suit against him, and then to have prosecuted me as his owner. As it is, he has instituted suit against me, but accuses my slave. This the laws do not permit. For whoever instituted suit against the master, and charged the facts against his slave—as though the slave had any authority of his own?Footnote 118
In this case, the master could be found guilty in a legal action brought against his slave, and could acquit himself by delivering the slave to the opposing party, according to the law of noxal surrender (though this is not mentioned in Hyperides’s account). The procedure did not make the slave a legal subject, strictly speaking: rather, it distinguished two defendants by referring all liability to a single person, the master. The responsibility entrusted to the slave was therefore only passive, or more precisely a kind of negative imputation, since all of his actions were ultimately referred to his master.Footnote 119 This imputation, in which the slave served as a support, offered masters significant room for maneuver,Footnote 120 but in no way did it imply a limitation of their liability in regard to their slaves.
Slavery and Representation: Another History of Athenian Democracy
We have, I hope, seen what a comparatist approach, here understood in its morphological form, can bring to the in-depth study of a particular slave society. It is not just a question of resolving the specific problems posed by incomplete documentation: the Ottoman muhataba does not simply allow us to suggest more or less ingenious solutions to the questions posed by Demosthenes’s text on Pasion’s slaves. Above all, when it is based on a clearly defined (or “typified”) configuration, the comparatist approach makes it possible to reinterrogate all of the arrangements that characterize the institution of slavery within a society. If we take this a little further, we gain insight into the institutional construction that linked Athenian politics to slavery around the question of representation.
It is understood that the modern concept of political representation, in its various meanings, was the product of a new world order established in the last centuries of the Middle Ages.Footnote 121 The “medieval revolution of the interpreters,”Footnote 122 that is to say, the rediscovery of Roman law by civil and canon jurists, played a decisive role in the process. It has been shown, by Pierre Legendre and Laurent Mayali in particular, that the canonical doctrine of representation originated in the rules of private law derived from the Roman concept of the mandate, giving rise to a new and far-reaching “system of authority.”Footnote 123 In the medieval mandate, as conceived by Martinus, the link between proxy and principal became one of perfect representation, in that the rights acquired through the representative’s acts were directly imputed to the person being represented. Qui facit per alium est perinde ac si faciat per se ipsum: “He who acts through another acts himself,” asserted Boniface VIII in rule 72 of his Liber Sextus. The representative completely represents the person being represented: such was the great invention of canon and civil law during the Middle Ages. This differed considerably from the Roman mandate, under which the results of the act accomplished by the proxy were initially attributed to his own person, before he was obliged to transfer them to the principal. By the action of the mandate, the principal could claim accounts from his proxy if the proxy took too long to execute the transfer. Accordingly, in the words of Yan Thomas, “the representative was never a perfect substitute for the person who was represented, for the reality of his own existence stood as a screen between the two.”Footnote 124 In fact, it was only in the domestic sphere, in the context of the relationship that united the master to all those in his power (in potestate), including his slaves, that perfect representation was possible, for in a certain way it could only be realized through the negation of the representative as a legal subject.
In commercial matters, however, Roman law established an early notion of representation—which can be described as imperfect—via the relationship between a principal and his proxy and, in procedural terms, the realm of actiones exercitoria and institoria.Footnote 125 These measures effectively aimed to define the way in which the actions of a proxy could be imputed to a principal, by guaranteeing to any third party who entered into a contract with the agent the legal nature of the engagement. Even as they were attempting to institute a limited form of liability on the part of the principal—in addition, rather than in place of, that of his proxy—these measures also laid the early foundations for a form of representation. Through this set of procedures, aimed at protecting both the master and the third party, an unprecedented autonomy was accorded to the slave in terms of property, notably through the acknowledgment of his liability for the value of the peculium.Footnote 126 In the institoria system, the link between principal and agent was superimposed on the traditional relationship of dependence that united master and slave, to the extent that it undermined its very foundation: the negation of the slave as a subject of law.Footnote 127 In this respect, it is possible that the development of the notion of representation, regardless of its imperfection, dismantled the principle that made the slave a pure extension of his master.
We know enough to turn once more to slavery in Classical Athens, now with a very broad question to answer. Here, unlike imperial Rome or Ottoman society during the classical period, the master’s liability was unlimited. It is subsequently tempting to link this organization back to the broader absence of any form of representation which lay at the very foundation of the Athenian democratic regime. In Classical Athens, the bond between the people and their magistrates was in no way a bond of representation in the modern sense of the word: magistrates did not receive any mandate; as full members of the dēmos, they always remained under its supervision and could be held personally liable for actions committed in the exercise of their function. In the legal sphere, there was no rule organizing an indirect, or mediated, representation of parties or witnesses, whose attendance at the hearing was essential. Each party was supposed to speak in their own defense or make their own accusations. A citizen who was absent from the city could only testify in a specific form, with his testimony being written down and then transferred to a free man who was present at the hearing and was held legally responsible to the extent that he could be charged with bearing false testimony (pseudomarturia).Footnote 128 In this respect, a refusal of representation was at the heart of the democratic anthropology of Classical Athens.
The comparatist project makes it possible to shed new light on the unique institutional construction that, in Classical Athens, associated slavery with the refusal of representation and even with democracy itself. I do not claim to have established a mechanical connection that would intrinsically link the absence of representation to the existence of slavery. More precisely, my hypothesis is that the development of the institution of slavery, in the specific forms it took in fifth-century BCE Athens, prevented the development of the very concept of representation. By studying the role of public slaves (dēmosioi) in the Classical city, I have elsewhere attempted to show how the use of slaves attests to a resistance, on the part of the civic community, to the emergence of a concept of the state as separate from society. By making those who were charged with its administration invisible, the city warded off the danger of a state that could constitute itself as an autonomous agency, thus endeavoring to keep every form of power within the immanence of its own existence as a community. I have advanced the idea that the recourse to public slaves made it possible to avoid all forms of political representation, which the civic community could conceive of only in terms of dispossession.Footnote 129
The analysis of legal forms that organized the liability of a master in relation to the actions of his slave makes it possible to extend this hypothesis, shedding new light on the question. The absence of any kind of representation between two free individuals seems to be intrinsically linked to the full representation that bound slaves to their masters and implied their negation as legal subjects. Though slavery was the setting for the de facto accomplishment of perfect representation, its very existence hindered the de jure realization of all forms of representation between free individuals, since action through the mediation of another could only occur at the cost of their dispossession. In other words, the regime of perfect representation that was achieved in the master-slave relationship had as its corollary the refusal of all forms of representation that could connect free men, and it was only through the disintegration of the institution of slavery that legal forms of perfect representation between free men could emerge. In this respect, the institution of slavery was inseparable from the Athenian democratic experience.
