Mesopotamian Slavery

Abstract

A survey of Mesopotamian slavery provides an opportunity to evaluate how foundational or institutional the practice was in its earliest known forms. The entry to and exit from slavery were relatively asystematic, the status often ended up being temporary, and the functions and status of slaves are difficult to distinguish from those of other persons (e.g., servants and dependents) who also performed labor under coercion. This chapter considers these aspects as well as developments and changes in Mesopotamian slavery in the Old Babylonian period (ca. 2000–1600 BCE), focusing on this specific historical period as a model to comment on slavery in cuneiform culture as a whole (which lasted from ca. 3300 to 300 BCE).

“What are a slave-woman’s dreams? What are a servant’s prayers?”—Sumerian proverb

Introduction

This chapter introduces us to Mesopotamian slavery by focusing on the Old Babylonian period (hereafter, “OB”). This periodization is bounded both temporally (ca. 2000–1600 BCE) and geographically: Babylonia corresponds roughly to southern Iraq from Baghdad to the Persian Gulf, as opposed to the Assyrian north near present-day Mosul/Kirkuk. The OB is a historical time and place about twice as long and large as Classical Greece, with thousands of documents available to inform us about our subject. It is necessary to focus on one period, because to describe a single “Mesopotamian slavery” for the whole of cuneiform culture (from ca. 3300 to 300 BCE) would require us to concatenate evidence from at least ten major historical periods, and only result in a Frankenstein image of our subject.¹ I will therefore explain the situation in one time and place which typifies the subject, but also try to give a sense of where, when, and how specific features of slavery in other periods conformed to or differed from it.

OB slavery was primarily a mechanism to provide credit in an emerging mercantile economy—which otherwise had few if any stable financial institutions—rather than to satisfy labor needs in any other productive sector.² It was primarily in this way that Mesopotamian slavery had globalizing effects: merchants, slaves, and transacted values moved through local, regional, and supra-regional trading networks: trafficked slaves were moved to what is now southern Iraq from places of origin as much as 1000 km away which are now in Turkey, Syria, and Iran. This has implications for language and cultural contact, but the most durable consequence was the transmission of the socio-economic practice of using people as a form of capital. The trade in slaves began to establish homologies of prices, systems of credit, legal standards, and even concepts of personhood throughout the Near East.

The OB period was the age of Hammurabi, a time characterized politically by competition and conflict between many small territorial states; socially, by an emerging class of private householders and merchants; and economically by credit markets and great trading networks connecting Mesopotamia to the Mediterranean, Anatolia, Elam, and the Persian Gulf, with contacts extending to Oman and the Indus Valley. It is a period with a well-rounded range of sources for slavery, with evidence from law codes and lawsuits, letters and contracts, proverbs and bequests. These diverse text types all have their own conventions and limitations, but they allow us to build a portrait of slavery as a social, legal, and economic institution and even of the experiences of individual slaves and masters.³

These records reveal the basic socio-economic contours of OB slavery.⁴ Some slaves were born into status (i.e., “house-born”); others were brought from foreign places.⁵ Some slaves eventually “integrated” into their masters’ families by adoption, inheritance, or marriage; some formed families of their own. Slaves could own moveable property; some were literate; they had the capacity to give legal testimony.⁶ Households which owned slaves generally had only one or two. Slaves and masters knew each other face-to-face as individuals, and slaves were virtually always referred to by name.⁷ They performed housework and field labor, but no type of work which was not also done by non-slaves. They were trusted with a wide range of tasks which required their free movement (even unguarded travel to other cities), independent judgment, and personal knowledge of their masters’ affairs. They acted as legal and business proxies for their masters, and their reputations reflected on their households. Yet these positive indications of personhood are tempered by other features. Slaves were bought and sold as chattel much like oxen and commodities. They were commonly used as pawns to establish commercial credit, and a variety of mechanisms existed to control their movement and behavior, including community surveillance and force. Slaves were required to wear distinctive hairstyles and perhaps garments, partially isolating them from their previous social identities, and they were “genealogically isolated” by the limited possibilities to pass on their names as patronyms, though this was not forbidden.

This mixture of adverse and positive attributes, broadly accurate for slavery in all periods of Mesopotamian history, requires us to ask where to inscribe the boundaries of “social death.”⁸ The social death which characterizes slavery cross-culturally, according to Orlando Patterson, entails the loss of name and identity, punishment by death for flight,⁹ and the rejection of the slave as a member of society. But Babylonian slavery, in its particular admixture of social recognitions and subordinations, does not meet all these criteria.

It must be stressed that we have no clear semantic or etymological understanding of the various Sumerian and Akkadian words for “slave” (either masculine or feminine) and “slavery.” This is a complicated matter which cannot be addressed in full here but suffice it to say that although some proposed meanings are hardly impossible, they are neither linguistically clear nor as logical as they first appear.¹⁰ For instance, Sumerian arad₂ and Akkadian ardu for “slave” might connote foreignness on etymological and even graphemic grounds (as “mountain man” or “(one) brought down (the river?).” But the etymons are neither certain nor even semantically exclusive in either case, and even the logical premises are complicated by the fact that both words were routinely applied to “slaves” who were clearly not foreign at all but (former) citizens of the same city. We can also exclude the idea that racial or ethnic identity was imbued in the Mesopotamian terms, as “Slav” is within “slave.” The point is that any translation of “slave” for Mesopotamian terms (including as I use it here) is purely conventional, and possibly anachronistic and plain old wrong.

All of this introduces us to the central problem and paradox of Mesopotamian slavery: the disposition of persons whom we identify in socio-economic terms as “slaves” because they were legally and commercially chattel, but who nevertheless had both agency and specific social and legal identities. Thus, even the evidence from this earliest of literate world cultures requires us to question how foundational slavery really was and the degree to which it compromised or erased personhood. Should we understand this Babylonian slavery to be para-social, in which a degree of personhood was conferred on slaves, at least to the minimum extent necessary to deputize them as effective proxies for their masters? Or should we think of that very construction as slavery’s most fundamental hypocrisy, to create a class of persons who could be “trusted” to perform as many roles as free persons without recognizing and permitting basic control over their bodies, places of residence, and life choices? No less urgent are the questions for the slaveholding society itself: how did the institution, practices, and individual slaves affect what it meant to be a Babylonian who was not enslaved?

Entering Slavery

When and how did a person become a slave in OB Mesopotamia? The discussion below moves from dispreferring certain options for mode of enslavement—namely, as consequences of legal punishment, raiding, or commercial markets—in favor of debt as the broad and underlying condition for most instances.

Probably no one, to begin with, was enslaved as a punishment by law in practice. A half dozen OB law collections (such as Hammurabi’s) regulated various matters regarding enslaved persons, but only rarely ventured to form their status.¹¹ The OB law collections we have, preserving around 450 paradigmatic decisions, were primarily concerned with masters and commercial practice rather than the regulation of slavery as such. These laws addressed five basic areas of practice:

• The award of damages to a master for the death or injury of a slave.

• Holding slaves as pledges or pawns in distraint for credit (see below).

• Limiting (but not excluding) slaves’ capacity to make commercial transactions.

• Punishment of flight, but largely focusing on the free persons who aided fleeing slaves, not the slaves themselves; the crime was understood to be a theft of property by a free person, not a slave’s decision to flee.

• The status and inheritance rights of children born of slaves and free persons. The laws mostly guaranteed the eventual manumission of such children, but sometimes only following the death of the enslaved parent. The law codes differed in supporting, limiting, or nullifying the rights of these children to inheritance of a free father’s property.

But all of these laws focus on slaves already in status. Only three of ca. 450 extant OB laws can even potentially be read as creating slave status. The first (SLEx 4’) says that an adoptive son who repudiates his family may be “sold for his full value,” though it does not say “into slavery.” The second and third are both found in Hammurabi’s laws. ¶141 states that a wife found at fault in a divorce case must reside in her husband’s house “as (if she were?) a slave,” though it is unclear why it is qualified “as (if)” (kīma); ¶146 says that a slave woman who bears her master’s child and then “aspires to equal status” with his primary wife cannot be sold by her (presumably out of jealousy), but should “counted (immanūši) among the slave-women”—which is of course confusing, because she was already a slave to begin with. It is possible that these decisions created slave status as legal punishments, but none of the three is crystal-clear on this point. Neither is any of them documented in any real-life context, which raises considerable doubt that these rules were ever applied.¹² No other crimes or civil infractions were said by the statutory laws to be punishable by slavery, and no court cases reflect such a punishment. The laws instead were overwhelmingly concerned with rules for people who were already masters and slaves to guarantee property rights, not to create new slaves.

Significantly, when the laws mention in passing any specific mechanism of status formation, it is sale.¹³ This accords with the fact that the most substantial body of evidence we have for enslavement comes from slave sale contracts. These contracts are the documents in which people are first (to our knowledge) called “slave” in their life histories. The slaves sold in these texts are people who are not previously or subsequently known to us in any non-slave status; no sale contract makes mention of any different and preceding status or cause of enslavement. It is therefore almost impossible to reconstruct the events by which any individual was sold let alone what happened to them next. Whether or not sale per se was the mechanism which created slave status cannot therefore be determined—especially given that many sales of slaves were demonstrably re-sales. (Other significant modes of transfer included gifts, dowries, and inheritances, but this was not entry into status per se.) It is hard to avoid the conclusion that slavery was defined as a contractual and commercial matter between private parties and not institutionally generated. It is worth noting, given the discussion of self-sale and enslavement for debt below, that sold slaves, although always named by the contracts, were never parties to the sale, although sometimes their family members were.

There is little evidence that OB slavery was sourced from the widespread capture of people through war raiding. That is, although there is occasional evidence for raiding for people subsequently forced to work for their captors—a form of unfree labor likely a distinction without a difference for those thus exploited—none of it points to the sale of those prisoners as chattel slaves.¹⁴ Indeed, prisoners of war were typically redeemable through the payment of a ransom (ipṭēru) by their home country’s family, temple, or palace. That OB slavery was not substantially stocked with prisoners of war may surprise readers who are familiar with the ample record of the mass deportation of conquered populations by many Mesopotamian states since the early third millennium.¹⁵ But even during the time of the Neo-Assyrian empire, when literally millions of people deported during three centuries of war violence, the evidence for the transformation of war captives, plentiful though they were, into slaves is “rather limited,” “difficult if not impossible to trace,” and “questionable.”¹⁶

Even when OB kingdoms captured prisoners of war, the actual work to which they were put is not well documented. The only evidence for this in the OB comes from an archive of texts from Uruk covering about eighteen months over the years 1742–1740, documenting prisoners of war mostly from local cities (not faraway places). The archive largely documents the individual households to which these men were assigned and their food rations, but says almost nothing about their labor. We need not doubt they did work—they were called “workers” rather than “slaves,” and none were sold—but there is almost no documentary attention to the kinds or amounts of work they did.

In fact, the capture and deportation of prisoners are topoi entirely absent from the royal literature we would expect to mention them, which tracks with the near-absence of slaves in mass-use by institutions. Palace and temple administrative texts, documenting properties, production, and people, are in good amounts throughout the four centuries of the OB, but we do not find much evidence that they depended on enslaved workforces. We do not, however, have from Babylonia the kinds of coherent royal archives such as the one recovered from the contemporary Syrian city of Mari (ca. 400 km northwest of Babylon), where slaves disbursed to private households could be summoned for seasonal mass labor in the king’s orchards and smaller numbers of female captives manned textile workshops of the palace.¹⁷ It is possible that such arrangements existed in Babylonia proper, but none of the surviving letters or work rosters hint at it. On the rare occasions when groups of slaves do appear as laborers, they were working alongside non-slaves (soldiers, workers, tribesmen, neighbors, etc.) performing the same tasks.¹⁸ In sum, there is little evidence that OB states organized the forcible acquisition of people in large numbers for labor exploitation or sale.

Instead, the OB slavery represented in contracts and letters was overwhelmingly characterized by relations between small numbers of specific individuals known to each other by name. Sale contracts usually conveyed just one person, occasionally two (often a parent and child sold together); rarely, three. In letters from prospective buyers asking merchants to procure slaves for them, again, the typical request was for a single slave, although occasionally up to six are mentioned, but we never hear about the acquisition of slaves by the dozens or hundreds. The letters make clear that it was a normal expectation that slaves could be acquired on the market, sometimes in foreign places by long-distance merchants, but also that the supply was limited, or that only “unhealthy” or “skinny” slaves might be available. Sometimes a preference for certain foreign ethnicities was expressed, perhaps for linguistic reasons (i.e., to obtain a slave who at least spoke some Semitic language, as opposed to Elamite or Hurrian¹⁹) or because “house-born” (i.e., native) slaves might come encumbered with rights to eventual manumission or inheritance. But nothing suggests that enslavement was thought to be particularly appropriate to specific ethnicities; many ethnonyms attached to slaves also described free persons.²⁰ Neither do literary sources present a consistent point of view about slaves as ideally foreign or not. One may consider two proverbs offering conflicting advice, even though they derive from the same composition. The first warns against the acquisition of foreign slaves:

When you bring a slave girl from the hills, she brings both good and evil with her. The good is in the hands; the evil is in the heart. (Instructions of Šuruppak, ll. 193–197)

while another argues for the superiority of foreign slaves, likening house-born slaves to “herbs that make the stomach sick”:

You should rather bring down a foreign slave from the mountains, or you should bring somebody from a place where he is an alien. My son, then he will pour water for you where the sun rises, and he will walk before you. (ll. 158–159)

No clear view can be gained on this point, unless it is deep ambivalence.

In any event, physical force does not seem to have played an important role in the acquisition or transportation of slaves. Contracts and letters speak of slaves bought in the way that other commodities were bought, with no hint of violence: no chains, shackles, brands, yokes, ropes, guards, or other apparatus are mentioned. Restraints were sometimes used on slaves who had fled, and a few instances of corporal punishment are hinted at.

Letters (which can provide more dependable contexts because they were composed to deal with specific rather than paradigmatic problems) strongly indicate that enslavement was a common result from the default of private household debts (Sum. e₂ ur₅-ra, a “household under obligation”), where the sale of selves or family members was done in exigence, but voluntarily.²¹ The Babylonian concept of “debt slavery” had a fundamentally conditional character: persons held in this way were understood to be pledges (Akk. nipûtu, a “distraint”) for debts to be repaid—parked in the household of a creditor until (someday) redeemed—rather than property whose ownership was unconditionally and permanently transferred. All such “enslavements,” though indefinite, were not considered a permanent change of status. That the verb of transaction in contracts was unambiguously “to sell” (Sum. šam₂, the same verb used to sell a sheep or a chair) may be misleading to the extent that slaves could later be “redeemed” (Akk. paṭāru, uššuru, ṭarādu; see “exits,” below). To put this in culturally salient terms, “slavery” was often a matter of a low-ranking member of a poor household being deputed to go and live and work (but also to be provided for, i.e., fed and clothed) in another household until such time as a family debt was repaid.

This form of enslavement could hardly be called a desirable option for the enslaved person, and we ought not avoid casting a critical eye on the discursive paternalism which characterized and apologized for it; for instance, the common Babylonian encomium to “take care” of slaves also carried with it the sense of “to keep an eye on” them. But like many other invidious choices, enslavement as a pledge was a rational economic alternative for seller, buyer, and sometimes even slave. It may have been preferable, in hardship conditions, to have your son or sister eat poorly as a slave in a creditor’s household for a year or two than to have them starve in your own. These choices were made within a cultural logic in which all household members—selves, wives, children, servants, some animals, and even real property—were not only social beings but also valuable and saleable according to ideas of proportional value, expressed in legal and even mathematical terms. To this extent, the social violence of slavery was not exclusively focused on slaves; the threat of enslavement hung over most non-slaves, who could in theory be sold any day, usually by the decision of someone else. The unpalatable choice of enslavement was created by a world in which commercial and financial economies were still emergent: one where long-distance, local, and financial markets existed, but sources of credit and financial instruments were few. As an economic phenomenon, Babylonian enslavement was what happened when the inequalities of household hierarchies were commodified for a marketplace which had few other financial structures to build on.

From the perspective of the creditor/enslaver, taking a person as a pledge was a mechanism meant to guarantee the repayment of a loan, most often extended for productive purposes, but it also entailed costs of feeding and housing someone. That the ideal was for distraint-for-debt to be temporary is exemplified by one of Hammurabi’s laws, which states:

If a debt is outstanding against a man and he sells or gives into debt service his wife, his son, or his daughter, they shall perform service in the house of their buyer or of the one who holds them in debt service for three years; their release shall be secured in the fourth year. (¶117) ²²

The way things were supposed to work was: for the term of a loan, a distrained slave was sustained in exchange for the same labor they might normally do in their original household; repayment was the hoped-for outcome on the part of all parties involved. That this did not always work out this way in practice is addressed below.

Debt-and-default was also the primary motor for the enslavement of foreign people. That some slaves were indeed foreign has disguised the fact that they were not acquired by raiding but were already enslaved when purchased in distant non-Babylonian cities. An analysis of sale contracts shows that, just like “native”-born slaves, they too had most likely become enslaved because of economic hardships in their households.²³ The majority were women, perhaps unmarried or widowed, some with children in tow, or even infants at the breast; this demographic profile suggests that women were often the “least valuable” members of poor families precipitating out of their households in order to alleviate the debts of their fathers, husbands, and brothers. The sale of individuals in contracts may also obscure that they document only single members of what were entire households collapsing under debt, with family members sold one by one and pushed out of local communities, on to market towns, and thence to Babylonia.

This image is preserved for us in one literary composition referring to “the banished enemy, the slave from the mountains, or the laborer with a poor wife and small children comes, bound with his rope of one cubit….” The likelihood that debt was the major stimulus for enslavement is furthered by the last part of the second proverb quoted earlier, recommending the acquisition of foreign slaves; it goes on to count among the benefits of having a foreign slave that:

he does not belong to any family, so he does not want to go to his family; he does not belong to any city, so he does not want to go to his city.

In practical terms, debt slavery for re-sold foreign slaves was substantially more permanent than it ever was for Babylonian ones, even those born into slavery domestically (wilid bītim, “house-born”).²⁴ Brought hundreds of miles from the home which had sold them in the first place, who would (ever) come to redeem these foreign people? Perhaps slavery was intended only to have been a temporary status—the short-term solution to a very specific problem—but the accretion of foreign slaves into the system may have created (mutatis mutandis) a more durable class of unfree persons over time because they were sold at a distance. The redemption of foreign slaves occupied a prominent position in Hammurabi’s laws: the antepenultimate and penultimate legal decisions were concerned with facilitating the return of persons purchased as slaves in foreign lands, pointing to frequent abuses of the redemption system (itself guaranteed by law). This may hint that permanent enslavement was emerging as a problem wanting a solution.

We may make an analogy to modern-day abuses where foreign guest-workers, made liable for outlandish transit and housing fees by coercive employment contracts, are effectively transformed into unfree workers, e.g., Filipino laborers in Kuwait or Latin American migrant workers in the United States. In systems of social inequality, the economic nature of exploitation is most directly elucidated by identifying the means of specification used to create, regulate, and maintain that inequality. The value of Babylonian slaves was commodified by law and contract; their status as pledges specified by a debt theoretically equivalent to the value of a human, redeemable but sometimes unpayable; and the gradual transformation of foreign slaves into permanent chattels suggests that a system originally based more on human trafficking for credit may have moved toward the reation of permanent chattels. The transfer of the debt obligation or the sale of the slave to a different geographic region could then prevent redemption, even if the original term of unfree service was finite; effectively, the system may have begun to create a category of permanently enslaved people. These developments presented new practical and even moral problems in a system meant to afford only temporary expediences.

The Labor and Economic Function of Slaves

Here, I distinguish between two topics: the labor of slaves and the economic function of slavery. The reader may reasonably expect that a report on the work of Babylonian slaves is the natural answer to the implicit question of why people bought and kept them. We assume that labor was (and is) the raison d’être for slavery—that slaves were kept to grind grain, mine silver, and tend the house. Were they not? Why else would people bother to buy, feed, and clothe slaves if not to profit from their labor?

But the subject of slave work is oddly subdued in the cuneiform record. Sale contracts never specified any particular skills of slaves being sold or work to be done; hire contracts (since slaves could also be hired out on wages basis) only occasionally did so, usually grain-grinding and weaving for women and fieldwork for men. I am aware of only a single OB letter specifying skill as a criterion for purchase: “As for the slave you wrote me about, buy her if she’s house-born and a weaver” (AbB VI 4). Stray passages in literary sources reflect ambivalent stereotypes of slave labor, both negative (that slaves were lazy, unreliable, fearful, grumbling, misbehaving) and positive (capable, honest, tireless). Compare, for instance, these two proverbs:

A free weaver equals two slave girls. A free worker equals three slaves.

At harvest time, at the most priceless time, gather like a slave girl, eat like a queen!²⁵

The first proverb implies that slave labor was inherently less productive; the second implies diligence. But parsing these brief and disparate characterizations is less helpful than observing how few literary references there are to slaves’ work at all.

OB letters paint the same picture. Of 804 references to slaves and servants in OB letters, only 6 percent were about work expected or performed. (By comparison, 37 percent of all statements had to do with where, when, and how slaves were to move freely to different locations to carry and acquire information, and 19 percent related to issues of pledge and distraint.²⁶) Most references to labor were unspecific—just that slaves should “do work,” “be on duty,” or “not be idle.” A few letters speak of herding, cooking, fetching wood, household chores, or field- or canal-work (including as proxies for labor expected of their masters); six letters mention specialized skills: weaving, gardening, grooming. But no form of work was particular to slaves: the work they did included the same tasks performed by free persons: harvesting, grinding grain, tending animals, dredging canals, etc. No activity was socially marked in the OB or any other period as distinctively “slave labor.”

What is more revealing is how typical it was for slaves to work alone, very occasionally in twos and threes, and virtually never in labor gangs. The few households that owned slaves rarely seem to have had more than one or two at a time. Slaves almost always worked without supervision and alongside non-slave persons. Force and coercion are rarely indicated (and even then, not uniquely for slaves). And though the dress and distinctive hairstyles of slaves visibly identified their status publicly, chains and other restraints were used only as forms of punishment in unusual circumstances; branding, used for slaves in the mid-third and mid-first millennium,²⁷ is not attested for the OB.

If we put these observations in dialogue with the foregoing discussion about entry into slavery, the somewhat counterintuitive takeaway is that OB slavery did not exist in order to aggregate labor power, solve labor shortages, or extract work in exchange for invested value. Slavery was not really about work, per se. To be clear: all slaves worked, but only as all other people did. Like most pre-modern agricultural societies, work was not an optional activity for anyone—even the highest priests and officials were obliged to perform corvée labor, for instance—and there was little opportunity to intensify production or maximize profits. To think that labor in this socio-economic setting could be commodified and apportioned to an encumbered class of persons so that another class could thereby enjoy leisure time is not only anachronistic, but just simply not in evidence. Moreover, given the technological and informational boundaries of the period, the raw compulsion of labor would have been wildly inefficient. Given the dispersed distribution of slaves throughout the population, forced work would have required supervision at a ratio too close to 1:1 to effect compliance. Neither did slavery solve any obvious deficits in available labor resources, given that other forms of semi-free labor existed alongside it: servants (ṣuḫarū), dependents (rēdûtu, i.e., in the status of a client; miqtu, a “fallen” person), tenants (ikkararātu), hirees (agrū), levied workers (tupšikkū, rendering ilku-duty), subordinates (mārū, lit. “sons”), and dozens of other terms for various kinds of workers. Finally, ganged labor under force and supervision would have hampered the most common task we see given to OB slaves: to carry messages, gather news, act as proxies for their masters, and be deputed as pledges for productive loans. In economic terms, slavery existed because it had these specific functions related to information and credit, not to labor power.

Free movement was central to these functions. The business of merchants and landowners required that they be in more than one place at one time, and few communications technologies were available to solve them. To check on the plowing of a field, collect a debt, deliver a letter, bring silver to a creditor, herd sheep, or ready a departing caravan, Babylonian masters needed slaves carrying letters to effect business when and where they themselves could not attend. Whether they were dispatched to other households, out to the countryside, or even to other cities altogether, it was precisely for their ability to move freely and exercise independent judgment that slaves were needed. For this reason, masters prized slaves not for brawn or beauty, but for being “trustworthy” (taklu), and socio-economic relations based on trust are fundamentally incompatible with those based on force. If we assume that “work” should look like hard labor under a beating sun, our expectations are confounded by the economic functions slaves fulfilled in a mercantile economy focused on commerce and credit.

The Experience of Household Slavery

Above, I have made two partially contradictory statements: that it is possible to say something about the individual experiences of slaves and masters; but also that we usually only have a historical window through single documents (usually sale contracts), which do not allow us to reconstruct life histories.²⁸ But a few vignettes and exceptions to the one-document rule allow us to illuminate social experiences of slavery in markets, worksites, households, and communities.

To begin from some admittedly crude statistics: the OB urban slave population (and slavery seems an emphatically urban institution) hovered around 5 percent; the average number of slaves we see in a household is one; an average household size was around five non-slave persons. Incidental evidence suggests that more slaves were women than men, perhaps by as much as a 2:1 ratio.²⁹ Given these parameters, we might find one slave in every fourth household along a city street, and perhaps 500 slaves in a Mesopotamian city with a population of ca. 10,000 people. Slaves were therefore common and familiar (but not typical) members of urban communities. This image of small-scale household slavery may differ from, e.g., institutional workhouses known from the third millennium, where large groups of foreign men, women, and children, some branded in their flesh, toiled in such places as a textile factory called the é kešda, the “Bound-House.”³⁰ We know that OB slaves were visually identifiable by a shaved hairstyle called an abbuttu-lock, perhaps wore distinctive garments, and were routinely watched and observed (especially when in proximity to city-gates, through which they were forbidden to pass). It is enough to observe that Mesopotamian cities in the OB were not particularly large, and that there was a great deal of face-to-face recognition and knowledge of people (including slaves) by name. If the shackles and racial identification we might think of as classic markers and constraints of slave life were not therefore features of this particular historical context, slaves were nevertheless under watch through pervasive community and individual knowledge—through the watchful nosiness of village life. (And slaves were themselves participants in social surveillance, sometimes reporting on the doings of neighbors and collecting juicy gossip.)

Contracts identified some slaves as “house-born,” which acknowledged them as natally Babylonian. Some laws promised eventual manumission and sometimes inheritance rights for such children born to slave women, and sometimes for the mothers as well.³¹ To this limited extent, our sources suggest that legal and social distinctions were made between slaves who had been born into households and those who had not. Foreign slaves, meantime, as far as we can make out, came from lands across a vast arc of the Near East, stretching almost 2000 km from the region of modern Aleppo in the northwest to Khuzestan in the southeast. The linguistic or cultural differences of these slaves from Babylonians seem to have been unimportant, however, to any documented dimension of their daily working life; foreignness was noted only in contexts of procurement, sale, and law, aspects related entirely to issues of property and jurisdiction. In no other respect is any preference or dispreference for any particular kind of slave expressed in any source beyond stipulations about health.

We do know that slaves could (and some did) marry, have children, live together with their families in their own residences, own moveable property (but not land or houses), and work fields to produce their own crops. How common these dimensions of life were remains entirely unknown. Slaves could marry into the families of their masters or take on a new status as caretakers for the elderly, as second wives, or even as adoptees. In such cases, they might in legal and social terms become freer and fuller social beings with a patronym and household identity. Some slaves were owned within the community of unmarried religious votaresses called nadītus, who conducted financial business as independent householders, but it is unclear that life in this community differed in terms of (m)any social or economic norms.

Whether “integrated” into masters’ families or not, slaves represented those households to the community, and their behavior and reputation, good or bad, “whether honored or despised” (as AbB III 11 puts it), reflected on their masters. A slave’s reputation went hand-in-hand with the “trust” (taklu) vested by a master to carry out household business. Slaves also had limited legal capacity, for instance, to give testimony though not to bring suit, but their legal liability was unequal, since they could be punished fully for various crimes (although it was their masters who were often liable for civil penalties and fines).

But slaves could exacerbate tensions within a household as well as reflect its identity. Squabbles between siblings often emerged about who should inherit a favored slave or whether a manumitted slave had received too much inheritance themselves. As with any family, membership hardly resolved all social, financial, or even emotional problems; and although manumission might improve a freed slave’s life chances in social and economic terms, upward mobility was severely limited in this society even for free persons. From the point of view of a slave, life in a household probably represented safety against true immiseration—starvation and homelessness, but “emancipation” was hardly any protection against the kind of debt which could push them right back down into slavery. There was much to lose, and not so much to win; unfortunately, such outcomes can rarely be tracked.

The evidence for structural violence in OB slavery is unclear. Notwithstanding the existence of children born to slave women, we have close to zero evidence about sexual predation. At most, we find oblique statements in letters like this rebuke: “With me, you talk about slave-women!” (AbB IX 19: 11–12); it is hardly clear what this means. Sexual abuse may have been so common that it went unremarked upon; it may have been entirely unusual or forbidden; we simply do not know. It is worth commenting only that namritu, an adjective commonly describing female slaves and sometimes translated as “good-looking” (therefore suggesting their acquisition for sexual use) actually meant only “healthy.”³² The evidence for physical violence is similarly scant; a very few references imply the use of beatings to compel obedience, but not to any degree more than for the corporal punishment of wives, children, or annoying neighbors. Again, this does not by itself tell us whether violence was common, everyday, and unremarkable—or highly unusual and shocking. Imprisonment and deprivation are in evidence here and there, as with a female servant reporting that “cold and lice are eating me” in the house where she was distrained (AbB VIII 100), but again the typicality of such complaints is unclear. All we can say is that open references to sexual abuse, physical violence, and deprivation are few; nor even idioms or allegories connecting slaves to tropes of sexuality or punishment.

If we look to literary sources for insights on social view of slavery, we find a pervasive ambivalence. Slaves appear quite frequently in Sumerian literature of the OB period; more than a hundred references can be found in dozens of works. But the scribes composing and copying proverbs, hymns, and stories were not interested in making sociological observations; they narratively positioned masters and slaves in symbolic roles to allegorize issues of authority and position, especially to satirize haughty behavior. Reversals were common devices, as in the proverb: “When the mistress left the house and the slave woman came in from the street, away from her mistress the slave woman set up her own banquet.”³³ But no consistent viewpoint emerges: some proverbs depict slaves as unfortunates (“Left-over clothes are the share of the slave-woman’s child” [SP 4.43]), while others show them as greedy and overfed (“Although the chickpea-flour of the home-born slaves is mixed with honey and ghee, there is no end to their lamentations” [SP 1.47]). One could say that they reflect an implicitly critical view of slavery insofar as it produced people who were either pitiable or contemptible; but the divergent archetypes reveal a profound ambivalence, seemingly to the purpose of explaining (or explaining away) the moral dissonance of slaves’ simultaneous subjugation and personhood.

Exiting Slavery

If Babylonian slavery was not by design meant to be an unalterable status, legal terms of enslavement were nevertheless indefinite and could therefore sometimes be made effectively permanent. That one might remain a slave indefinitely is demonstrated by the fact that slaves were awarded as damages in lawsuits and considered heritable, as shown by inheritance texts (ṭuppāt aplūti), dowries, and laws. That slaves were distinguished (or at least listed separately) in inheritance texts from “possessions” (Akk. būšu, i.e., mere “things”) does little to ameliorate that they were chattels whose possession could be transferred—by gift, sale, default, or bequest. It seems likely that any time a slave was transferred from one household to another—and therefore to a new master—the term of their enslavement and/or how they might be redeemed were both rendered less clear, not more. We may compare this to the more permanent status of foreign slaves branded as temple or royal property in the mid-third millennium; or the unfree temple oblates called širkū in the mid-first millennium, who (since they were not slaves as such) “could be neither manumitted nor sold.”³⁴

We do read that some slaves ran away—but not from slavery, only to be slaves in other households (sometimes nearby, sometimes in other cities) which they preferred, and which presumably preferred them. Sometimes distrainers kept possession of slaves in violation of contracts and/or against the will of the slaves, but choice of residence by slaves played a remarkably common role in resolving household conflicts. The Babylonian idiom, in fact, was not that slaves had “run away,” but merely “wandered” (ḫalqu³⁵) from where they were meant to be, as a stray ox might; there seems no expectation on anyone’s part that slaves were attempting any wholesale escape from status. To the extent that the problem was policed (i.e., not much in this period), this was aimed squarely at the recovery of the master’s property, not the punishment of the slave.³⁶ None of this precludes, as I have said, that slaves were watched, marked, and controlled—only that there was not much of anywhere to run away to. That being the case, restraint and arrest were exceptional measures used in an environment where the mobility of slaves was required.

Most commonly we read of slaves being “redeemed.” This practice of pawning a slave for credit and then redeeming him/her when profits were realized (on the borrower’s sale of barley or the return of his caravan) is attested many times. There were a number of ways to say this: often with verbs denoting the satisfaction of debts (e.g., Akk. paṭāru, wuššuru), but sometimes only descriptively, i.e., that slaves were “returned,” “sent back,” or “handed over” (Akk. târu, ṭarādu, nadānu), where the reason is unstated. In many instances, the original debt had been repaid, and the creditor returned the distrained slave to the original master. We know of slaves who were sold and re-sold in this way multiple times between different masters, probably among circles of merchants who were all in debt to one another in turn. But often we read about distrainers not returning a pawned slave, despite being paid or going past the normal term of distraint. This is where the law stepped in: if a slave or a master could bring proof that a creditor had either been compensated or that the original debt was somehow invalid.³⁷ It is clear, however, that a slave “redeemed” from a debt was not freed from slavery, only from distraint, because they were returned to the household of an original or earlier master.

Finally, we may contrast redemption with emancipation: certain OB laws mandated the “release” (andurāru³⁸) of slaves or prohibited “claims of slavery” against individuals (e.g., ana wardūtim ul iraggum³⁹), usually for slave women who bore children to a master. In such cases, a full manumission was to be effected, involving a full exit out of slave status. (There were, however, limits on the liability of the master to recognize the inheritance rights of the children and/or the marital status of the mother.⁴⁰) What is less clear is how often this actually happened outside of the idealizing logic of the laws: most texts about manumissions using these terms are only model contracts used by scribes as learning tools; we have few actual practice texts. A few documents describe manumissions in terms of ritual actions, such as a “purification” (Akk. elēlu, including with pūtu 1a-2´, “to purify the forehead”⁴¹) or a declaration before a divine image; in other contexts, these actions were used to indicate the clearance of debts, which suggests that manumission was conceived of in terms of financial solvency. The smashing of a pot at the feet of the slave (Akk. šebēru 1b), anointing with oil (Akk. šamnu 1f + šapāku), and the shaving-off of the abbuttu-hairlock are also attested.

Manumissions most frequently coincided with a transition to some other specific status. Many slaves being “freed” were simultaneously obliged by new contracts to support their former and aging masters as “sons” or “daughters”, to become votaries in temples, or to marry into the family.⁴² Few manumissions simply released slaves from servitude, but facilitated some new status with different obligations. The terms of the new arrangement seem typically to have improved the lot of the former slave—there was almost no way for things to get worse, after all—but we need to recognize it was usually a “freedom” which came at the price of a commitment to another obligation. The contrast between “free” and “slave” is, in fact, one made by astoundingly few texts, whether letters or contracts⁴³; manumissions did not contrast slavery to any other civil-legal status, such as amīlūtu or muškēnūtu, “free” or “semi-free.”⁴⁴ This was a society in which the paramount social good was to ensure that people were established somewhere within a household hierarchy, not to award (much less guarantee) anyone legal and political rights in any community larger than a household. That household membership (even as a slave) guaranteed some measure of socio-economic security, however, implied its opposite: the threat of exclusion (and thus possible starvation) was a form of structural violence which hung over all members. Just because we are not faced with a record full of beatings or sexual violence (which may anyway have been unnoted and normalized aspects of class and gender relations) does not mean that a very substantial and even existential threat of ostracization did not lurk behind the façade of paternalism and compliance that characterized mastery and slavery.

Conclusion: Paradigm and Variation

The foregoing has described a particular form of slavery, one in which force, restraint, and labor extraction were relatively minor concerns, and reliability, agency, identity, and mobility were considered advantageous characteristics of a slave. Above all, slavery was geared toward mobilizing productive credit in an emerging mercantile economy based on moving agricultural goods to market; where the commodified value of slaves offered one of the few available financial instruments in the marketplace.

It is not my purpose to depict Babylonian slavery as inherently mild or humane, but to point to its economic function as based on social relations of trust. Given this, a determination of “social death” seems unwarranted; not all exploitation nullifies social identity. I mean to stress what the institution was for rather than how it compared to other historical instantiations, e.g., the mass-labor context of racialized slavery of the antebellum US South. The more urgent question about Babylonian slavery, then, is not about the degree of exploitation of the slave, but the larger consequences for families and communities in which all people were potentially salable—where every mother, daughter, wife, younger brother, or son could be pledged against a bargeful of barley. Perhaps they would be redeemed next week, perhaps in three years, perhaps never. We have few sources directly commenting on the social and emotional stresses this system must have put on Babylonian households, but it would be obtuse not to consider it as a larger context. It reminds us that enduring structures of inequality are not only harmful to the people who directly suffer deprivation and exploitation, but also to the people who supposedly benefit but could also fall prey to them—for whom legacies of privilege and moral compromise entail profound psychosocial harm. Systems of inequality are harmful to all who participate in them.

How well does the OB situation (ca. 2000–1600 BCE) reflect slavery in earlier and later periods of Mesopotamian history? This can be difficult to answer; other periods produced documentation substantially different in quantity and kind. Corpora prior to the OB, for instance, preserve large numbers of administrative texts from specific times and places—Uruk ca. 3200, Lagaš ca. 2500, Nippur ca. 2300—but with several centuries remaining effectively silent, and with few contracts and no letters. This situation begins to improve around fifty years before the OB begins (i.e., around 2050) under the Third Dynasty of Ur, when commercial documents become more abundant and letters begin to appear. The documentary record following the OB, meanwhile, is relatively sparse from ca. 1600–750, after which point it picks up momentum and widens into a vast river, down to the fourth century, producing texts of all kinds—letters, contracts, laws, and literature. The three thousand years of Mesopotamian history thus (very!) briefly profiled are not known to us on a consistent documentary basis in temporal, spatial, or generic terms, which makes comparisons difficult; the available evidence may artificially emphasize or elide different aspects of slavery.

What we can do is note large-type differences in period documentation, if not necessarily in practice. We may begin by noting that the term(s) for “slave(ry)” develop somewhat later than the institution’s apparent existence. The term “sag + arad₂/geme₂” from the Uruk period (3300–2900) does not so clearly designate unfree laborers . It is not until the Akkadian period (2350–2200) that we have our first evidence for the sale of slaves by contract (a commercial practice extending across the Near East over time)⁴⁵; only by the Ur III period (ca. 2112–2004) do we see slaves given as pledges for loans. During these earlier periods (and later ones as well), not only did other words for unfree persons persist alongside slaves, but terms distinguishing them from “free” persons were relatively undeveloped. The OB introduced the concept of a “gentleman” or “free” class (awīlū) but it remained inchoate in juridical and social terms. By Middle Babylonian times, a privileged status for city-dwellers existed (kidinnūtu, referring to divine protection), though this substantially meant tax exemptions rather than privileges over other classes. Only by around 700 was there a clearly marked free class called mār banî (lit., “grown son”) openly contrasted to slave status. The legal and social differences between slaves and non-slaves is therefore not clear to us until relatively late times. This complicates our thinking, and invites us to think about why a distinction between slavery and freedom was not made more systematically than it was.

Another possible major difference between OB and third-millennium slavery is the question of whether, how often, and how many slaves were obtained through capture in war. There remains a low-level and sustained (but perhaps irreconcilable) debate about whether the (undoubted) trains of prisoners brought back to Mesopotamia from foreign wars and set to work in large groups on temple and palace estates (as described by Early Dynastic and Akkadian kings) were saleable chattel. The massed POW laborers (perhaps even stationed in labor camps) and the individual slaves of private households, though both victims of force, may have formed distinct sub-populations, casualties of substantially different systems of inequality.⁴⁶ We should, of course, stop to ask how relevant our modern distinctions about status would have been to a captured Subarean or Elamite soldier in, e.g., 2250 BCE: roped around the neck, marched hundreds of miles away from home to the strange land of Sumer, branded in his flesh, and forced to work on a temple estate or gifted to a local grandee by the king, he probably cared little what his precise juridical status was or whether the people who oversaw his work called him a “slave” or not. We can only say that slavery in subsequent times, beginning with the Ur III period, took on a more emphatically financial and commercial character, with little war capture or massed labor.

A more subtle difference can be seen in the Neo-Babylonian period. The use of slaves as pledges for loans had remained a long-term feature of Babylonian commerce, but the enslavement of free persons for debt became effectively illegal after ca. 700, effectively limiting distraint to persons already in slave or subordinate status. Further, manumissions and other exits from slavery seem to have become rarer and more conditional, and permanent marking practices such as branding now obtained (in contrast to the special haircut of the OB, which could obviously be undone). These changes may indicate that the greater specification of the category “free” (mār banî) may have come at the price of a hardening of the category “unfree,” with less social mobility (either upward or downward).

But Babylonian slavery largely remained a stable set of practices. In all periods, we see the enslavement of both Babylonians and foreign persons, and of men, women, and children. Slavery remained a highly individualized arena of interaction between masters and slaves known to each other by name, identity, and character. After the third millennium, there are fewer instances of unfree labor at scale, for any kind of labor being particular to slaves, or for slaves as creating any outsized economic value in either primary or secondary production. Despite a possible increased stratification between slave and free persons in the latest periods, slavery remained unmarked in moral or racial terms; little suggests that slavery was ever considered natural to any particular (set of) person(s).

So I briefly return to the very difficult questions posed at the outset of this essay about slavery’s implications for the personhood of everyone involved in this culture—sellers, buyers, creditors, debtors, merchants, judges, witnesses, neighbors, bystanders, and, of course, slaves. On the one hand, one could emphasize the human scale and temporary nature of Babylonian slavery as less harmful than more dehumanizing instantiations of the institution; we do not find the routine violations of human rights and brutality characteristic of some other historical cases. The close context of household life, an array of expressions of care and affect, relations of trust, and a seemingly high incidence of mixed free and slave marriages all suggest some degree of personhood for slaves. The need to use slaves as an extension of a master’s authority suggests that violence would undermine their economic purpose. Full “social death” cannot describe this form of slavery, even as contradictions and double standards persisted (as they do in all cultures). To this extent, Babylonian slavery looks “nicer.”

On the other hand, Babylonia could be said to be responsible for a generative and standardizing role in the globalization of slavery, since it stands at the opening of a very long, continuous, literate, and self-conscious historical tradition. We have to wonder about the role of the writing system in particular, almost obsessively organized to create categories of meaning, to mark for the first time the social worth of certain persons in relation to others; writing itself was a means of specification. It is not difficult to see writing’s categorical imperatives extending first to the practices of war and commerce that ultimately connected individual Babylonian households with far-flung parts of the world; then to a regularization of the financial means of specification for debt which translated categories of personhood into commodities; and, finally, the hardening of these terms into legal rules and social institutions. The apparently lower levels of violence and inequality visible in the Babylonian record for slavery was not a product of moral imperatives, but of insufficiencies of financial and technological means: this version of slavery was not a matter of being “nice,” just of a social engineering calibrated to a specific economic end. If this culture had had the capacity to intensify production and control people at scale, it would have been perfectly happy to do it. If only for normalizing slavery, the Babylonian case was indeed “foundational” to the globalization of the institution throughout the pre-modern world—out to other ancient Near Eastern cultures, which adopted many Mesopotamian standards of price, status, sale formulae, and legal practice—hence to Greece, Rome, the rising Islamic world, medieval Europe, and beyond.

Notes

• 1. All dates used in this article other than bibliographic contexts are BCE. The major periods of Babylonian history include: Uruk, ca. 3300–2900; Early Dynastic, ca. 2900–2350; Akkad, 2334–2192; Ur III, 2112–2004; Old Babylonian, ca. 2017–1595; Kassite, ca. 1475–1154; Middle Babylonian, 1154–ca. 750; Neo-Babylonian, ca. 750–539; Persian, 539–331; Seleucid, 331–63. “Sum.” refers to terms in Sumerian; “Akk.” to terms in Akkadian. Translations of literary works follow the Electronic Text Corpus of Sumerian Literature (ETCSL: https://etcsl.orinst.ox.ac.uk); references to laws follow Martha Roth’s Law Collections from Mesopotamia and Asia Minor (Atlanta: Scholars Press, 1995); translations of Akkadian words, The Assyrian Dictionary of the University of Chicago (CAD) (Chicago: The Oriental Institute, 1964–2010). The abbreviation AbB refers to the series Altbabylonische Briefe in Umschrift und Übersetzung (Leiden: Brill, 1964–2005).

• 2. This is supported by studies of other Mesopotamian periods as well: see, e.g., Heather Baker, “Degrees of Freedom: Slavery in Mid-First Millennium BC Babylonia,” World Archaeology 33, no. 1 (2001): 23, reporting: “The tablets rarely give us information as to what duties the slaves performed”; and Seth Richardson, “Walking Capital: The Economic Function and Social Location of Babylonian Servitude,” Journal of Global Slavery 4, no. 3 (2019): n. 15 for literature. Cf. Vitali Bartash “Going for the Subarean Brand: The Import of Labor in Early Babylonia,” Journal of Near Eastern Studies 77, no. 2 (2018): 263–78. Bartash argues for the economic significance of slave labor in Sargonic Babylonia.

• 3. I prefer “master” to “owner” partly because it inclines to the semantics of Akkadian bēlu, but also for the argument made below that Mesopotamian debt slavery typically positioned the slave as a pledge, where possession was not indentical to ownership (if conditionally).

• 4. See especially Richardson, “Walking Capital,” 285–342.

• 5. See F. Van Koppen, “Geography of the Slave Trade,” in Mesopotamian Dark Age, ed. H. Hunger and R. Pruzsinszky (Vienna: Österreichischen Akademie der Wissenschaften, 2004); Seth Richardson, “Origin of Foreign Slaves in the Late Babylonian Period,” KASKAL 17 (2020): 53–73.

• 6. Slave women are attested as witnesses to legal documents as early as the Akkadian period.

• 7. Slave names are rarely given with a patronym, however, which denoted full and recognized social identity. As for the re-naming of slaves, the evidence is mixed: slave names of both undoubted foreign origin and probable “re-christenings” are attested in many periods. But re-naming does not appear to have been a consistent practice. For exceptions in the Neo-Babylonian period, see Baker, “Degrees of Freedom,” 22.

• 8. Orlando Patterson characterized slavery in many (if not all) world cultures as “social death” in these terms. See Orlando Patterson, Slavery and Social Death (Cambridge, MA: Harvard University Press, 1982). Baker engages with Patterson’s ideas in her study of Neo-Assyrian slavery (Heather Baker, “Slavery and Personhood,” in On Human Bondage, eds. J. Bodel and W. Scheidel [Malden, MA: Wiley-Blackwell, 2017]).

• 9. Even OB “laws” at most punished slave flight by cutting off an ear rather than inflicting death. Two points may be made, however: one, we have little indication of how often these punishments were actually practiced; two, Hammurabi’s law to punish a slave rejecting his master (¶282, the last in his list) was probably a literary metaphor for the punishment of political rebellion and not the enforcement of slavery per se.

• 10. See further my article in prep, “Mesopotamian Words for ‘Slave’: Opacity and Mutability in Early Terms and Practices.”

• 11. The “laws” found in compositions like the Code of Hammurabi were not statutes as such, but collections of the notional and paradigmatic decisions of just rulers. That I occasionally refer to them by shorthand as “laws” is simply to avoid annoying the reader.

• 12. One may compare with the Neo-Babylonian period, when enslavement is documented as a punishment for, e.g., infidelity (Baker, “Degrees of Freedom,” 21).

• 13. With the Akk. verb šâmu, “to sell,” distinct from agāru, “to hire,” including the hire of labor.

• 14. See esp. Richardson, “Origin of Foreign Slaves”; cf. Jacob J. de Ridder “Slavery in Old Assyrian Documents,” in Kültepe International Meetings, Vol. II, eds. F. Kulakoğlu and G. Barjamovic (Turnhout: Brepols, 2017), 56f.

• 15. Notably by the Neo-Assyrian (934–612) and Neo-Babylonian (626–539) empires. There is substantial evidence for the capture and deportation of prisoners of war from Early Dynastic and Akkadian inscriptions. Instances in which those prisoners are called “slaves,” however, are few.

• 16. Baker, “Slavery and Personhood,” 22; see also Ignace Gelb’s classic study “Prisoners of War in Early Mesopotamia,” Journal of Near Eastern Studies 32 (1973): 70–98: “Is it possible to turn POWs into slaves? The answer is theoretically, yes, practically, no!”

• 17. I thank Hervé Reculeau for this observation.

• 18. On the poor evidence for POWs/captives as slaves, the absence of mass/institutional contexts, and the emphasis on individual social relations, see Richardson, “Walking Capital,” 3 n. 9, 9 with n. 28, 16–17, and 23; and Richardson, “Origin of Foreign Slaves,” 55–57.

• 19. See Bartash “Going for the Subarean Brand,” 263.

• 20. “Elamite,” for instance, was an ethnonym identifying both slaves and free land-owning citizens. Similarly, V. Bartash (“Going for the Subarean Brand,” 273–74) notes that although most “Subareans” were slaves, some were merely menial laborers.

• 21. This practice is attested from as early as Ur III times; see S. J. Garfinkle, “Shepherds, Merchants, and Credit,” JESHO 47, no. 1 (2004): 5, 7 n. 16, and 27.

• 22. See again Garfinkle, “Shepherds, Merchants, and Credit,” 9 and 27, writing of the Ur III situation: “This does not mean that customary loans could not result in debt slavery, but only that this was not the primary intent of the creditor in all such arrangements.”

• 23. Richardson, “Origin of Foreign Slaves.”

• 24. Cf. the Laws of Hammurabi, ¶¶117–18, which specify that slaves (of any origin) given into debt service could be kept or sold by the creditor if the debt was not paid within three years, whereas non-slaves in the same position were to be released.

• 25. SP 3.183 and “The Instructions of Šuruppak” ll. 131–133.

• 26. Richardson, “Walking Capital,” 305–23.

• 27. Bartash “Going for the Subarean Brand,” passim; Baker, “Degrees of Freedom,” 22.

• 28. Cf. Bartash “Going for the Subarean Brand,” 269–70, on the (brief) life history of a third-millennium slave woman named “Goody.”

• 29. See, e.g., Bartash “Going for the Subarean Brand.”

• 30. Bartash “Going for the Subarean Brand.”

• 31. Cf. Baker, “Slavery and Personhood,” 17–18, who is less sanguine about apparent Mesopotamian protections against “natal alienation,” at least for first-millennium contexts.

• 32. On the material conditions of OB slave life, see Richardson, “Walking Capital,” passim.

• 33. Perhaps the most famous composition based on this kind of reversal/mirroring rhetoric is the later Akkadian composition “The Dialogue of Pessimism,” in which a slave mechanically endorses a series of contradictory actions proposed by his master, satirizing his vacillation.

• 34. Bartash “Going for the Subarean Brand,” 275; Baker, “Slavery and Personhood,” 24. Baker notes that manumission is simply not attested in Neo-Assyrian texts at all.

• 35. Cf. the less common Akk. abātu, more straightforwardly “to flee.”

• 36. LE ¶¶33–34, “defraud”; LE ¶¶49–52; LH ¶¶15–16.

• 37. In which case penalties for the faulty distrainer might apply, including if the slave had been injured or died while in their custody: LL ¶14 (two-fold); SLHF viii 11–15 (equal amount); LE ¶¶22–23; LH ¶280.

• 38. See the note in CAD A/2 s.v. andurāru s. (p. 117) on the term’s use for both release from slavery and cancelation of debt. The Sumerian term for andurāru (ama.ar.gi₄) means a “reversion to a previous state,” which may imply a presumption that free status preceded slavery.

• 39. See also CAD Q s.v. qabû 3a-1´b´ and qerēbu 5a.

• 40. LL ¶25 (slave mother and children freed; no inheritance share); LL ¶26 (if the master  marries a slave after his first wife’s death, child gets inheritance share); LH ¶119 (with paṭāru), ¶¶170–71 (the latter with andurāru), ¶175; cf. LH ¶176a–b. More ambiguous is LH gap ¶s, which seems to suggest that a slave whose master beats him need not be returned to him, but it is unclear if the slave is actually freed.

• 41. This may indicate that the forehead was where the abbuttu-lock of hair was placed: see Kraus Texte 25 cited s.v. abbuttu in CAD A/1 p. 49, “if he has a low growth of hair on his forehead (pūtišu) as far as his abbuttu…”; cf. SLHF ii 4–13.

• 42. See exs. s.v. CAD Š/3 širku A s. and šikūtu s. (esp. A32117), devotion to a temple, and šitektu s., given in adoption; CAD Q qerû v. 2b; palāḫu v. 5g, marriage; ḫarrānu 9, with an adoption.

• 43. Cf. H. Baker, “Degrees of Freedom,” World Archaeology 33, no. 1 (2001): 21: Neo-Babylonian slave sale contracts clarified that a sold person was (also) not a free person; such distinctions were not made in the OB.

• 44. Cf. the slightly later situation at the nearby city of Nuzi, where manumission could be qualified as a transfer to a semi-free status termed šalaššu. On concepts of freedom, see Eva von Dassow, “Freedom in Ancient Near Eastern Societies,” in The Oxford Handbook of Cuneiform Culture, eds. K. Radner and E. Robson (Oxford: Oxford University Press, 2011), 205–24.

• 45. Slave sale contracts are next found in Elam in the twenty-first century; in Assyria, by the nineteenth century at Aššur and then in its colony in Kaneš; in Alalakh by the fifteenth century; in Emar by the fourteenth century; etc.

• 46. This distinction is probably even more valid for the case of war captives in the much later imperial Neo-Assyrian period, when large-scale deportations mostly had the character of mass resettlement rather than mass enslavement.