Slavery, Freedom Suits, and Legal Praxis in the Ottoman Empire, ca. 1590–1710

Abstract Beginning with the story of the Muslim youth Mehmed bin Abdülcelil of Tunis, this article examines the plight of Ottoman subjects abducted and sold into slavery within the Ottoman Empire and their efforts to regain freedom through Ottoman courts. Freedom suits (hürriyet davaları) were common in the seventeenth-century Ottoman Empire, so much so that contemporary legal praxis manuals (sukuk) always provided examples of how to document them, but they have never been systematically studied for this period in which slave ownership was extremely widespread and the legality of enslavement depended solely on religion and subjecthood. Drawing on a sample of seventy-nine suits from greater Istanbul and eleven sukuk manuscripts, this article considers how the trade in the illegally enslaved was concealed by the immense traffic in licit captives and how the theoretical protections of Ottoman subjecthood clashed with the practical challenges of how to prove it, exposing the gap between slavery as legal institution and slaving in practice. Whereas the vast majority of freedom suits ended in rulings in favor of the victims, most of the illegally enslaved probably never managed to have their cases heard or were turned away for lack of evidence.

Knights of St. John imprisoned hundreds of Muslim captives awaiting sale or ransom. 1 Ironically, Mehmed's fate was to fall into captivity not on Malta, but in rural Anatolia. Just a few stages into the journey, near the northwestern city of Eskişehir, the adolescent Mehmed was abducted by two men who marked him as a fugitive slave and brought him to market in the village of Gecek. Terrified of his captors and fearing further abuse, Mehmed answered in the affirmative when a potential buyer, Allahverdi bin Musa, asked if the men were his owners, and he encouraged Allahverdi to buy him. Convinced, Allahverdi bought Mehmed for a mule and two large silver coins (guruş, here riyal guruş, i.e., Spanish reals). Mehmed soon fled the unsuspecting Allahverdi, thereby becoming what his abductors had falsely represented him as: a runaway slave, whose name and description-"open-browed, dark-eyed, swarthy (esmer), of average height"-were circulated in the surrounding jurisdictions among the court officers and fugitive hunters. Eventually apprehended, Mehmed was brought before the local authorities, who summoned Allahverdi to retrieve his property. But in judicial custody Mehmed was able to speak freely: He was not a slave at all, but a Muslim "of free origin" (hurrü'l-asl). On 25 October 1644, one year after his abduction, Mehmed and Allahverdi faced off in Istanbul before the highest-ranked judicial figure in the Ottoman Empire, the military judge (kazasker) of Rumeli. 2 In the chambers of the Imperial Council (divan-ı hümayun) in the second court of Topkapı Palace, the administrative heart of the Ottoman Empire, five Tunisian notables testified: "The aforementioned Mehmed was born in the Bab el-Cezire neighborhood of the city of Tunis and his parents are of free origin," they said. "His father, Abdülcelil, is free, and his mother, Fatima bint Ali, is free and thus the aforementioned Mehmed is free." It is unclear whether these men had traveled from Tunis expressly to testify or had been part of Mehmed's pilgrimage group or already resided in Istanbul, but on the strength of their testimony the court declared Mehmed legally free.
Yet Allahverdi could also summon witnesses, and these confirmed that Mehmed had said "I'm a slave, buy me" (ben kulum beni al ) in the village market. The jarringly direct, utterly unambiguous words attributed to Mehmed meant that he had "deceived" Allahverdi, who now swore an oath that he did not know Mehmed was "of free origin." The result was that Allahverdi was not criminally or financially liable. Since Mehmed's kidnappers were nowhere to be found, responsibility for compensating Allahverdi fell squarely on Mehmed's young shoulders instead.
And so, a third evidentiary phase of the proceedings commenced to determine how much Mehmed would owe. As Allahverdi had paid with a mule, two expert witnesses familiar with the animal in question were summoned for their appraisals. Although Allahverdi had claimed that it was worth 98 guruş-for a total purchase price of 100 guruş-these witnesses declared that, in their expert opinion, the 1 aforementioned mule was worth no more than 70 guruş. The judge therefore ordered Mehmed to pay Allahverdi 72 guruş, the revised value of the mule plus two guruş in coin. 3 Though a bargain compared to ransoms on Malta, 4 it was still a significant sum. Just as redemption was beyond the means of most Muslims brought to Malta, many in Mehmed's position would be forced to negotiate service-forpayment plans with their captors that amounted to years of continued enslavement in all but name.
Mehmed bin Abdülcelil's freedom suit in the court of the Rumeli kazasker was unusual in some respects. With its three evidentiary phases the record is longer than most, taking up an entire folio page, and it is written in two different hands, suggesting a lag between the initial complaint and the final evidentiary phase of the proceedings. But from the exalted venue down to the phraseology of the record, the case bore much in common with the many suits heard by Ottoman judges (kadis) of the seventeenth century from plaintiffs claiming to be "of free origin." Like the vast majority preserved in the records, it was successful, and it hinged on the in-person testimony of male, Muslim witnesses. 5 "The basic principle (as : l) among men is freedom ( _ hurriyya)," wrote the Damascus-based Hanbali jurist Ibn Qudāmah (d. 1223), relating the universally accepted view in classical Islamic jurisprudence that slavery was an exceptional condition and presumptive freedom the rule whenever in doubt. 6 In a legal opinion (fatwā), the celebrated Ottoman chief jurisconsult Ebu Su'ud Efendi (d. 1574) concurred: To the question of whether a man should be released if he denied that he was a slave and his captor could not prove otherwise, Ebu Su'ud declared, "Yes, with his oath, [since] the basic principle (asl ) in the Abode of Islam is freedom (hürriyet)." 7 But this principle of "original freedom" (hürriyetü'l-asl) was still subject to the rules of Ottoman-Islamic legal procedure. In court, the burden of proof lay not with the captor-the defendant-but with the captive-the plaintiff. That burden often proved unbearable for society's most vulnerable.
What, then, did it mean to be "of free origin" in a massive, multiethnic empire with porous borders but without government identity papers or passports? If a male Muslim from a well-off Tunisian family was at risk of capture and enslavement less than 300 kilometers from Istanbul, what dangers faced the less well-connected, especially women and younger children? Drawing on little-known manuscript sources and a sample of seventy-nine freedom suits heard in Istanbul-area courts between 1590 and 1710 (H. 999-1121), this article explores the concept of "original freedom" underpinning these cases, then establishes that complaints of illegal enslavement were common, difficult to prove, and subject to powerful socioeconomic forces and increasingly standardized record-keeping practices that slotted them into predictable, preexisting templates of legal documentation, or else 3 Ibid. The record does not specify what language(s) Mehmed spoke; I have transliterated his name according to Turkish rather than Arabic convention in keeping with the language of the record. 4  concealed them forever. 8 In this era of widespread slave-holding and unmet market demand, cases like Mehmed's illustrate the limits of the power of juridical labels and categories-not least "slave" and "free"-that were often mutable, concealable, or ignored. * * * * * * Freedom suits are an unparalleled source for understanding how slavery was defined, legally justified, and socially constructed. The declarations of litigants and the decisions of judges help us to locate the social and legal boundaries between "free" and "slave"-which in the Ottoman Empire and in the wider Islamic tradition were essentially defined in opposition to one another 9 -and the paths and pitfalls between the two. Freedom suits illuminate the gaps between slavery as institution and slavery, and slaving, as historical practice. 10 They also offer historians the tantalizing prospect of recovering enslaved litigants' voices, underscoring their agency in often surprisingly successful efforts to navigate fundamentally unjust legal cultures. For all these reasons, scholars of slave-holding societies in the Atlantic world have written extensively on the legal efforts of the enslaved to win their freedom.
Across legal cultures and empires, from the sixteenth through the nineteenth centuries, plaintiffs litigating for liberty drew from the same quiver of arguments. These included asserting prior or promised manumission, descent from a free or freed woman (or man, in the Islamic context), abduction and illegal (re-)enslavement, and a sojourn (or origin) in "free soil" or territory. 11 Variations of all these arguments were made by enslaved plaintiffs in the Ottoman Empire, but a striking parallel for cases like Mehmed bin Abdülcelil's can be found in early nineteenth-century Portuguesecontrolled West Central Africa, where freeborn African plaintiffs alleging illegal 8 Suits for freedom were not uncommon outside these dates, but earlier sixteenth-century suits were typically recorded by scribes as modified Arabic manumission documents (which are far less informative) instead of litigations, while suits after the 1710s represent a shrinking, less diverse enslaved population. See Table 1 for the seventy-nine suits, which are unevenly distributed across thirty-two court registers from nine courts: Istanbul, Bab (a subsidiary of Istanbul's main court), Galata, Üsküdar, RSM, and one register each from Ahi Çelebi, Beşiktaş, Eyüb, and Hasköy. Roughly forty registers, divided equally between Galata and RSM, were consulted in person at İSAM, contributing sixteen suits to the sample. Forty registers were published online by İSAM and another sixty registers (nine of which were within the date parameters and contributed to this sample) by Medipol Üniversitesi Hukuk Fakültesi, all accessible at https:// kadisicilleri.istanbul. For Ahi Çelebi, see Feryal Örkmes, "1072 Tarihli Ahi Çelebi Mahkemeleri 11 Numaralı Şeriyye Sicil Defteri" (MA thesis, Marmara Üniversitesi, 2014). enslavement sued on the basis of their "original freedom" (liberdade de origem), a concept emerging from early modern Iberian jurisprudence that was predicated on Portuguese vassalage and/or Catholic faith. 12 In the Ottoman Empire, it was the litigations of plaintiffs declaring this sort of inherent freedom from enslavement that were called hürriyet davaları-freedom suits. 13 Scholars of the more-thoroughly-studied late Ottoman Empire have highlighted the efforts of enslaved Circassians and Africans to obtain their freedom in the second half of the nineteenth and early twentieth centuries via lawsuits in both the Islamic and new secular court systems, as well as through appeals to Ottoman administrative authorities and European, especially British, consuls. These struggles arose in the aftermath of the Ottoman Empire's promulgation of edicts proclaiming (male) equality under the law (1839, 1856), part of the Tanzimat program of reforms; the banning of the trade in enslaved Africans and shuttering of the official slave markets (1857); and Russia's forced expulsion of the Circassians in the 1850s and 1860s, which led to an enormous influx of refugees whose elites transformed Circassian hereditary serfdom into slavery in Ottoman domains. The persistence of slavery in this period of international abolitionism, even as the mechanisms for legal importation were banned, and the proliferation of venues where enslavement might be challenged led to the articulation of claims to freedom (hürriyet) on the basis of recent imperial legislation, constitutional ideals, or, in the case of many Circassians, to their being freeborn Muslims who were now subjects of the sultan. 14 In the last instance, petitioners were mobilizing legal arguments and language that had a long history in the Ottoman Empire, but in drastically changed political and ideological circumstances.
Slavery in the pre-nineteenth-century Ottoman Empire was not racially defined. Religion and subjecthood alone mediated the boundaries between those who could and could not be enslaved legally. Freedom suits heard in seventeenth-century Ottoman Islamic courts were thus contestations of both the legal status and the social condition of slavery that were grounded in an assertion of juridical belonging on the basis of Ottoman subjecthood, adherence to Sunni Islam, or both. The "original freedom" these plaintiffs claimed made no reference to ethnicity (cins), skin color (levn), or ancestral origin (asl), nor necessarily to the faith of their birth, but broadly speaking to Ottoman subjecthood, which free Sunni Muslims and voluntary foreign emigrant converts to Islam enjoyed by default and which otherwise applied to all freeborn inhabitants of the Ottoman Empire. 15 All who were not "of free origin" were slaves, former slaves, or theoretically enslaveable.
These freedom suits, though common in the Ottoman Empire, have never been studied in aggregate. 16 Ottomanists have exploited court registers (kadi sicilleri) for decades, fueling numerous studies on slave origins, occupations, escape and resistance, and manumission. 17 Yet for scholars focused on the quotidian activities of individual courts for relatively circumscribed periods of time, freedom suits like Mehmed bin Abdülcelil's-comparatively diffuse, seemingly aberrational-may have been easy to overlook or dismiss. 18 It is abundantly clear from the registers that manumission-a hallmark of Islamic slavery which theoretically facilitated the integration of former slaves into societydid not always go smoothly: masters reneged on promises of freedom or their heirs chose not to honor them, masters abrogated work-release contracts (mükâtebe) or reconsidered spontaneous declarations of emancipation. Even manumission agreements fulfilled to the letter might be violated years later by heirs seeking to boost their inheritance by incorporating emancipated slaves in the estate. And human traffickers routinely abducted former slaves, stole their manumission documents, and resold them some distance away. The resulting suits, filed by women and men who had at one point been legally enslaved, broadly resemble many of the litigations for liberty generated in the slave-holding regimes of the Americas wherein plaintiffs proved they had been freed or promised freedom. 19 But in pre-nineteenth-century Ottoman usage, the literal "freedom suits" were those like Mehmed bin Abdülcelil's, submitted by Ottoman subjects of all faiths and by foreign Muslims who had been subjected to the condition of slavery but who had been, legally speaking, "free" all along. Successful plaintiffs in these cases thus "proved freedom" (isbât al-hürriyet), even though they had lacked it in any meaningful sense when they filed suit and might not fully regain it afterward. 20 The number of suits of this type recorded in the Istanbul region likely ranged anywhere from high single digits to dozens per annum, and they were regularly heard in towns and cities across the empire. 21 Their significance is not a function of their quantity, however. As law and practice are most visible in the breach, asking how illegally enslaved plaintiffs described and disputed their condition; how Ottoman legal professionals heard, decided, and recorded their cases; and why suits succeeded or failed can offer valuable insights into the workings of Ottoman slavery and Ottoman subjecthood that are otherwise invisible in the legal sources upon which Ottomanist scholars rely. Freedom suits reveal not only how freedom and slavery were conceptualized in the premodern Ottoman Empire; they illuminate the dark spaces within this dichotomy where untold numbers toiled.

Sources
Examining freedom suits means examining court records. Identifying which ones to scrutinize, however, is no straightforward proposition, given the Ottoman legal system's jurisdictional flexibility and multitude of venues. Whereas proximity to the slave market may have determined where many freedom suits were heard-thus high concentrations in the Istanbul and Bab courts compared to the wider distribution of manumission-related complaints-the social status of the litigants and the nature of the litigation often funneled these cases towards the Rumeli Sadareti Mahkemesi (henceforth RSM), the court of the Rumeli kazasker. 22 The kazasker dealt primarily with the affairs of the askeri class (lit. "military," but encompassing everyone in imperial service), to which a significant proportion of Ottoman slave-holders belonged, and since defendants had the last word on venue, the kazasker heard many freedom suits with askeri defendants. 23 Additionally, the Rumeli kazasker was frequently tasked with adjudicating the claims of non-askeri petitioners to the Imperial Council and cases in which government officials had taken an interest. Lower courts often referred these cases to the kazasker, and informal Leonor Won Their Freedom: Litigation and Liberty in Seventeenth-Century Mexico," Hispanic American Historical Review 85, 1 (2005): 39-80. 20 White, "Slavery, Manumission, and Freedom Suits." 21 Approximately twelve courts operated in seventeenth-century greater Istanbul. Most registers cover roughly one year. Sixteen of the thirty-two registers supplying this sample (among which there is virtually no chronological overlap) contain three or more freedom suits; five registers contain two. 22 Founded in the early seventeenth century, the Istanbul slave market was adjacent to the Grand Bazaar; see Zübeyde Güneş Yağcı, "İstanbul Esir Pazarı," in Zübeyde Güneş Yağcı, Fırat Yaşa, and Dilek İnan, eds., Osmanlı Devleti'nde Kölelik: Ticaret, Esaret, Yaşam practice became official policy in the aftermath of the Greek Revolution, when out of concern for corruption the government ordered the kadis of greater Istanbul to transfer all freedom suits to the kazasker for adjudication at the Sublime Porte in the presence of the grand vizier. For all these reasons cases like Mehmed bin Abdülcelil's routinely appeared on the RSM's docket. 24 The ravages of war, worms, water, and fire have taken their toll on the court records of greater Istanbul, and there are significant gaps in the historical record for the sixteenth and seventeenth centuries. 25 These gaps, combined with scribes' tendency to set aside registers for certain notarial entries (e.g., probate and pious endowments) and return to older registers to fill blanks (to say nothing of later archivists' interventions), confound any effort to conduct random sample-based quantitative research. 26 While I offer broad observations on the basis of my sample -assembled from onsite archival work and registers published online-caution is warranted. Yet the standard alternative approach, an impressionistic qualitative study based on a handful of cases, or structured around a single extraordinary one, risks missing key patterns. Because Ottoman court record entries are brief and often inscrutable, it is essential to balance this evidence with other sources to identify the paths that brought plaintiffs to court and understand why they led to the recognition of freedom or return to servitude. To that end, we must look to Ottoman legal praxis manuals, known as sukuk, or sakk mecmuaları.
Sukuk proliferated in Ottoman Turkish around the turn of the seventeenth century, as the Ottoman legal establishment began to standardize the language and procedure of the courts outside the Arabic-speaking lands and phase out the use of Arabic. 27 Prepared by kadis or court scribes to facilitate this process and cement their legacies, sukuk provide examples of how to record dozens of different types of litigations and notarial transactions. Critically, they telegraphed the various scripts which plaintiffs' and defendants' statements and witnesses' testimonies were supposed to follow-or into which their actual words would subsequently be molded-depending on the facts of the case and the nature of the proceedings. It is from these sources, for example, that we can be certain that the declaration "I'm a slave, buy me" attributed to Mehmed bin Abdülcelil represented not his own words but a judicial cliché indicating a particular, predictable outcome, namely that he would be required to reimburse Allahverdi. 28 24 Twelve registers contributing to this sample are RSM, seven of which contain three or more freedom suits. For the orders to transfer freedom suits, see Istanbul 154, f. 26b (Evail/C/1240). 25  Aided by sukuk, scribes from across the empire's northern tier, from Albania to Anatolia, came to employ nearly identical phraseology, which took shape in the courts of Istanbul and other major cities and gradually filtered outward. These manuals represented the new, standardized Ottoman Turkish legal idiom to a broad audience of law students and legal professionals and ensured its replication, for the manuals were effectively "mad-libs" for judicial scribes, who could simply swap out names, places, and details as necessary. The examples given in sukuk were not hypotheticals. Authors used real cases, culled from their experiences, that were meant to be exemplary but not extraordinary, and so sukuk provide a useful framework for and counterbalance to research in court records, enabling us to draw conclusions about what was "normal" across a broad swath of the empire over time. Lawsuits contesting enslavement were so common in the Ottoman Empire that nearly every Ottoman Turkish judicial praxis manual produced between the late sixteenth and early nineteenth centuries contains multiple examples, including those filed by or on behalf of illegally enslaved Ottomans.
This article thus pairs its seventy-nine-suit sample (see Table 1) with research in eleven unpublished, virtually unstudied sukuk manuscripts. 29 Distinguished by the survival of (comparatively) abundant archival material in an increasingly standardized legal-administrative language, the Ottoman seventeenth century was marked by frequent disorder on the frontiers, regular upheavals in the capital that threatened centralized control, and costly, generation-long wars that failed to generate enough captives to meet the persistent demand for slaves from the most elite households to more middling ones and from a wide variety of industries. 30 It was, in short, a troubled period in which the opportunity and incentive to take Ottoman captives existed and persisted throughout the empire. Many of those captives would end up, sooner or later, in Istanbul.

Slaving in the Ottoman Empire
Slavery in Ottoman lands was regulated by Hanafi Islamic law and sultanic legislation and was practiced extensively, but the vast majority of captives were bought and sold in urban markets and most probably remained in urban settings. 31   manumission was common and slave status rarely inherited-the children of enslaved women and their holders were legally free-a constant stream of captives had to be imported to meet market demand. 32 Only "enemy infidels" (harbi kafirler), non-Muslims from outside Ottoman domains, could be enslaved legally. Free Muslims were supposedly off-limits, as were Ottoman-subject non-Muslims (zimmis), protected in exchange for their obedience and payment of an additional tax. Zimmis' protected status could be lost if they were seen to be in violation of the pact, such as by joining enemy forces, which transformed them into enslaveable "enemy infidels." The Ottoman state employed this logic repeatedly well into the nineteenth century to motivate the underpaid soldiery sent to quash rebellions in the European provinces with the promise of captives. 33 More legally problematic, the Ottoman state systematically enslaved zimmis in the Balkans and Anatolia through the infamous child-levy (devşirme), which supplied recruits for the janissary corps and the palace schools that produced governing elites, most extensively in the fifteenth and sixteenth centuries. 34 Nevertheless, providing justice and protection for his subjects, Muslim and non-Muslim alike, was the sultan's primary obligation in Ottoman political thought. 35 Through the first half of the sixteenth century, "enemy infidel" captives were obtained primarily through cross-border raiding and prisoner-taking in wartime, supplemented by Tatar slaving expeditions north of the Black Sea and, to a lesser extent, regular trades in slaves from the Caucasus and sub-Saharan Africa. As the scale and frequency of Ottoman military successes declined, however, the proportion of slaves on the Ottoman market supplied by the Tatars increased. Consequently, the vast majority of captives available in seventeenth-century Istanbul were "Rus'," most hailing from what is today Ukraine. 36 Demographics changed dramatically in the eighteenth century when traditional sources of slaves were cut off due to Russian expansion into the Black Sea region, the concomitant cessation of Tatar raiding (Ottoman support for which the 1700 Treaty of Constantinople prohibited), and the replacement of wartime captive-taking with a prisoner-of-war system. These shifts led to a steep decline in slave ownership and a much smaller enslaved population that, by 1800, hailed almost exclusively from East Africa and the Caucasus. "enemy infidel" was a capacious category theoretically including all non-Muslims in non-Muslim-ruled domains, the sources from which captives might be obtained legally were severely limited by the sultan's commercial and peace agreements with foreign powers. 38 These restrictions meant that Ottoman subjects, who unlike treaty-protected foreigners lacked diplomatic representatives in Istanbul to intercede on their behalf, could be tempting targets for unscrupulous opportunists. Court records and administrative documents preserve evidence of everything from large-scale human trafficking operations to one-off betrayals perpetrated by family, erstwhile friends, employers, and spouses. Pirates and paramilitaries ran networks that distributed captives far from their homes, often with the connivance of local officials. If removed a sufficient distance, such captives might be camouflaged by the surrounding slave population, and if they eventually proved their freedom in court, the perpetrators could rest easily knowing the consequences were unlikely to reach them; most culprits escaped punishment. 39 Individual soldiers returning from campaigns or garrison postings abducted women or children-the sight of soldiers with captives in tow being unremarkable in the European provinces-and sold them in major population centers. And whereas many perpetrators were socially marginal figures, it was evidently not uncommon for heads of households or their heirs to (attempt to) sell vulnerable, legally free domestic servants, often after many years of service. 40 While gender, age, appearance, and cins did not determine the underlying legality of enslavement, they had a marked impact on who might be saleable and at what price. Cins (Arabic: jins, derived from the Latin genus) in this period encompassed and often conflated what would be defined in today's English as race, ethnicity, and nationality, expressed through the Arabic compound "of ___ origin" (e.g., ifrenjiü'lasl, "of Frankish origin," habeşiü'l-asl, "of Abyssinian origin," çerkesü'l-asl, "of Circassian origin," etc.). 41 The size or specificity of cins categories and the meanings attached to them were not static over time and space, and peoples from places with which the Ottomans had more intensive relations, particularly Southeastern Europe and the Caucasus, were typically labeled with more precision, if not necessarily accuracy, than others. Since captives brought to Ottoman slave markets were slotted into one of the cins categories known to Ottoman scribes, illegally enslaved Ottoman subjects needed to be plausibly presentable as being of a cins in which enslavement was often licit and commercially desirable. Young women and boys commanded the highest prices, and were also, not coincidentally, those most likely to fall prey to human traffickers on both sides of the legal divide.
Returning to the case of Mehmed bin Abdülcelil of Tunis, we might reasonably ask what cins his kidnappers believed him to be, or be passable as, when they marked him as a fugitive. Since court scribes rarely noted cins for Ottoman subjects, who were above all "of free origin" (hurrü'l-asl), it is difficult to say. Some slave holders in more remote or unstable areas were not particularly worried about provenance or potential lawsuits-pirates and brigands in both Rumeli and Anatolia abducted locals whom they openly employed in shipbuilding, domestic and sexual service 42 -but for the buyers and dealers in major urban centers who did their business in governmentregulated bazaars and were guided by slave-buying manuals and physiognomy treatises, captives' provenance, cins, and general appearance played critical roles in decision-making alongside gender, age, and price. 43 Buyers and sellers who traded, even if unintentionally, in the illegally enslaved faced potentially enormous financial losses in the event of a freedom suit. That risk was not confined to the unregulated aftermarket, nor to those who neglected their due diligence. Frequent complaints about the sale of free people in the Istanbul slave market factored in the central government's decision to purge two-thirds of the slavedealers' guild's membership in 1640, though sales contracted outside the market by unlicensed dealers continued unabated. 44 It must be emphasized that the defendants in these suits were not unsophisticated dupes or socially marginal figures. They included high-ranking palace officials and senior provincial administrators, scholars, janissaries, kadis, and even an Ottoman princess, Gevherhan Sultan, daughter of Selim II (d. 1574). 45 To protect buyers, slave sale contracts throughout the medieval Mediterranean had often required the enslaved to acknowledge their status as slaves and "consent" to their sale, and this practice evidently continued, albeit irregularly, under the Ottomans. 46 A concern with legal provenance on the part of buyers and dealers might be motivated by more than risk avoidance, however. An Ottoman buyer morally comfortable with purchasing a Tatar-kidnapped Rus' captive may well have been disgusted by the prospect of Ottoman subjects being sold in the same space. 47 After all, the former scenario, subject to the state's regulation, was entirely legal, but to enslave a Muslim or a zimmi was unambiguously haram, a violation of God's law as much as the sultan's. 48 42 48 Multiple hadiths (sayings attributed to the Prophet Muhammad) identified the sellers of free people as the worst of sinners, and such sales caused considerable anxiety for authorities, Jonathan A. C. Brown, Slavery Ottoman subjects' vulnerabilities to illegal enslavement were geographically, politically, and militarily determined. Although Mehmed bin Abdülcelil's case is a reminder that any isolated youth was a potential target, communities located along the empire's frontiers east and west and astride well-trafficked military routes were at greatest risk, especially those whose inhabitants might be semi-convincingly conflated with enslaveable enemy infidels. Thus, Greeks from the Aegean islands were a tempting target in the aftermath of the conquest of Cyprus in 1570-1571 and during the lengthy maritime conflicts of the seventeenth century. Magyar-and Slavic-speaking Ottomans were frequently abducted, as were Christians, Muslims, and Jews from the Ottoman-Safavid borderlands in the Caucasus, particularly during the grinding wars of the late sixteenth and first half of the seventeenth centuries. Vlachs and Moldavians, Ottoman-subject residents of autonomous tributary states, were seized by Tatar raiders, by Ottoman forces intervening in the principalities' periodic revolts and succession crises, and by individual soldiers during tribute collection missions. 49 The Imperial Council consistently ordered that illegally enslaved Ottoman subjects be found, freed, and returned to their homes. 50 Rapid government intervention could be effective, but once captives were distributed to individual buyers, government agents were forced to work through the local courts. For captives from areas with limited Muslim settlement, overcoming the evidentiary barriers to success required determined effort from the government and the exploitation of legal loopholes, since only Muslims could testify when Muslim defendants denied their slaves were "of free origin." 51 As a result, suits lodged by Christians from isolated, predominantly Christian areas like the Aegean islandswhence abductions are well attested in administrative records-are exceptionally rare in the court registers and contemporary sukuk.
We encounter only one such case in our sample, and the plaintiff's convoluted journey to Istanbul is indicative of the persistence of illegal slaving in the Aegean and the reasons for its near invisibility in judicial records. Converted to Islam during her captivity, Hadice bint Abdullah sued a certain Mehmed Bey for her freedom in Galata, Istanbul's maritime hub, in 1687. Supposedly believing her to be Rus', Mehmed had purchased Hadice in Chania on Crete from a zimmi from the Ionian island of Lefkada, which had been reconquered by Venice in 1684. In fact, Hadice was from Karystos in southern Euboea, which unlike most Aegean islands had a substantial Ottoman administrative presence and a sizable Muslim population. A classic case of "slave laundering," pirates associated with the notorious Aya Mavra fortress on Lefkada had probably captured Hadice in an amphibious raid and then transferred her to a local Christian collaborator who exported her to Crete, where she was sold to Mehmed Bey. Since Muslims with ties to Euboea could be found in Istanbul-something that could not be said of the smaller islands-Hadice was able to secure admissible witnesses and won her release. 52 But there were undoubtedly many more like her who slipped through the cracks.
Although there are scattered references to the enslavement of individual Muslims in the late-sixteenth-and seventeenth-century registers of the Imperial Council's decrees, nearly all the orders to find and free groups of enslaved Ottoman subjects concerned Christians. 53 The court registers paint an entirely different picture, however, one primarily of Muslims, especially women, carried off singly or in small groups. In the sample of seventy-nine suits examined here, 62 percent were female (forty-nine), and of these women, at least 80 percent were Muslim (thirtynine), 14 percent were Christian (seven), and the remainder could not be identified with certainty. Notably, a significant majority of the women in the sample were Muslim-born. 54 Of the thirty men in the sample, 83 percent (twenty-five) were Muslim at the time of suit. At least seven of these were converts to Islam, but intriguingly three had converted prior to their capture, all of them from the borderlands, both east and west, whereas all the women who had converted to Islam did so during their captivity. The fact that many illegally enslaved Christians converted to Islam after capture (much like those legally enslaved) may be indicative of their youth, since those below the age of reason could be converted without consent and those only a little older could probably be induced to do so.
This then is not a simple story of the persecution of isolated and vulnerable non-Muslims. In sum, 85 percent of plaintiffs were Muslim when they sued for their freedom, and roughly half of the total were Muslim-born. Only about 13 percent of recorded plaintiffs were identifiably Christian. There is only one Jewish plaintiff in the sample, who originated in Meskheti (present-day Georgia), but there is additional evidence in the court records for the enslavement of Ottoman Jews over a wider imperial geography, particularly in the many suits lodged by former owners seeking refunds from sellers. 55 Ottoman illegal enslavement was ecumenical.
It was also widespread, and pervasive. More than half of the plaintiffs in the sample had their origins in the European provinces (especially present-day Albania, Bosnia, Bulgaria, and Hungary) and the tributary states of Moldavia and Wallachia; roughly 25 percent hailed from Eastern Anatolia, the Crimea, and Ottoman-held portions of the Caucasus. Whereas many called small villages home, quite a few were from large towns and cities, as well as communities residing inside major Ottoman fortresses. 56 One woman even hailed from Mecca. 57 While both the court 52 Galata 138, f. 59a (8/R/1098). On "slave laundering" and Aya Mavra, see White, "Piracy of the Ottoman Mediterranean." 53 Vatin, "Une Affaire Interne"; White, Piracy and Law, 36-58. 54 Of the thirty-nine women identifiable as Muslim, only 18 percent (seven) were definitely converts to Islam. Converts are sometimes explicitly described, while others can be identified by the adopted patronymic "ibn/bint Abdullah. records and administrative documents point to the frontiers as a major source for illegal captives, especially in times of conflict or major troop movements, only the court records reveal the extent of the internal threat: of the women in the sample, nearly 25 percent (twelve) were kidnapped and sold within greater Istanbul itself, and all but one of them were freeborn Muslims. Are these data representative of the contours of illegal enslavement in the seventeenth-century Ottoman Empire? No. For one thing, notwithstanding the diverse origins of the RSM's caseload, sources produced exclusively in Istanbul could never be wholly representative of the empire. But these data may be broadly representative of the circumstances of those who went, or were brought, to court and actually managed to have their cases adjudicated, both in and beyond Istanbul. Support for this assertion can be found in the sukuk (which also draw examples from litigation heard in Edirne, Bursa, and other towns and cities in the empire's northern tier). The eleven sukuk manuscripts surveyed contain a total of twenty-four examples of suits filed by free-origin plaintiffs and an additional eleven suits filed by former holders of the illegally enslaved seeking their money back from sellers. Approximately 30 percent of the freedom suits in the sukuk manuscripts were filed by Christians, and the remainder by Muslims (the proportions were similar in most individual volumes), about a third of whom could be clearly identified as converts to Islam (a few having freely converted prior to capture in scenarios mirrored in the sample). Thus, over half of the plaintiffs were Muslim-born in both the manuals and the registers themselves. It is tempting to assume those similarities are significant.
However, it seems unlikely that half of the Ottomans trafficked in this period were Muslim-born, given that contemporaneous imperial decrees to find and free the illegally enslaved predominantly concerned non-Muslims. Rather, the discrepancy suggests that Muslim-born plaintiffs are significantly overrepresented in the court records and manuals, thanks to the greater ease with which they were able to find witnesses and prove their cases, whereas Christians from areas with limited Muslim presence often had to rely on governmental intervention to secure release. Indeed, many freedom suits with Christian plaintiffs heard by the Rumeli kazasker bear the hallmarks of official intervention, including a hearing in the chambers of the Imperial Council, plaintiffs escorted by janissaries, and testimony from government officials or others whose personal connections to the plaintiffs appear tenuous.
It is in light of these facts that we must consider the astonishingly high success rate in these suits, 92 percent, with only six of the seventy-nine in the sample failing to "prove their freedom" in court. As for the manuals, only two-the late seventeenthcentury manual of Hızır bin Osman and the early eighteenth-century manual of Musazade, both rather lengthy and thorough-provide examples of failed freedom suits (one in the former, two in the latter). 58 The reason for the high success rate is simple: unsupportable claims rarely proceeded to formal litigation. The court registers do not record every complaint the kadi heard, only those that generated a document issued to the principals. Hopeless cases left no trace, making it impossible to estimate the actual numbers of illegally enslaved who came to court or what the success rate of "original freedom" claims would look like if adjusted to include petitions dismissed without litigation. Nevertheless, reading the court records and the manuals together enables us to recognize patterns and tell the stories of the enslaved while shedding light on how Ottoman subjecthood was defined and demonstrated in practice and the limitations of the protection it promised.

Suing for Freedom
Under the heading, "example of proving subjecthood (riayet) and free origin," the seventeenth-century Ottoman Turkish sukuk of Hızır bin Osman provides the following: The average [height], open-browed, blue-eyed bearer of this document, the Christian named Apnaku son of Fulan (i.e., so-and-so), and the tall, blue-eyed, closed-browed Christian named Alkisnedera daughter of Sumum, each sued Halil Beşe ibn Fulan and Ali Beşe ibn Fulan, who are present in court and holding them as slaves, and declared the following: "We are inhabitants of the village of Fulan (i.e., such-and-such) in the land of Moldavia, born of parents who are [Ottoman] subjects of free origin and we too are [Ottoman] subjects of free origin. Twenty days before the date of this document, when we were traveling from the aforementioned village to the town of Ismaila by cart to sell honey, we got stuck by the side of the road with a broken wheel, at which point the aforementioned Halil Beşe and Ali Beşe attacked us and tied us up and brought us to Istanbul. They are illegally holding us as slaves. Now we request that Halil Beşe and Ali Beşe be questioned, that their possession of us be invalidated, and that we be set at liberty.
Their claim to be Moldavian subjects (Boğdan reayasindan) was subsequently investigated in court and it became clear from their language that the aforementioned were Moldavian. When Halil Beşe and Ali Beşe were questioned, the aforementioned Halil Beşe responded, "about a month earlier in the town of Ismaila I bought the previously described Alkisnedera, a slave of Rus'ian origin, from someone named so-and-so and paid the price of one hundred guruş," and the aforementioned Ali Beşe responded "fifteen days ago in a place called Birgos Pina I bought the previously described Apnaku from someone named Silistreli Hasan and paid the price of eighty-one guruş by exchanging a packhorse of equivalent value," and in this manner they denied (inkar) the plaintiffs' claim that they are [Ottoman] subjects and of free origin. When the aforementioned plaintiffs were asked for proof of their claims, the barber el-Hac Dilaver bin Abdullah, who is originally from the lands of Moldavia but resides in protected Istanbul in the Tahta Minare neighborhood inside the Fener Gate, and Mehmed bin Abdullah, appeared in court to give testimony.
They testified that: "Indeed, the aforementioned plaintiffs Alkisnedera daughter of Sumum and Apnaku son of Kalu are from Moldavia and are residents of a village named Musku and were born of free-origin Ottomansubject parents and are Ottoman subjects and of free origin (riayet ve hurrü'lasldır). We are witnesses to this matter and we testify thus." After their testimonies were investigated their testimonies were accepted, in accordance with which the aforementioned plaintiffs were declared to be Ottoman subjects and of free origin. As a result of the legal decree Halil Beşe and Ali Beşe's possession [of them] was invalidated and the aforementioned were set free. What happened was recorded by request. 59 This example is fairly typical of how such entries, which took the form of the document (hüccet) issued to the litigants, were recorded in this period. A carryover from the medieval Arabic legal formulary tradition, one of the distinguishing features of Ottoman freedom suits is the inclusion of the stylized physical description (hilye) of the plaintiffs-height, eye and hair color, brow shape, sometimes skin tone and distinguishing marks-that appeared in all documents concerning slaves, including the manumission documents (itkname) that made them free. Whereas the example above concludes that the plaintiffs are "of free origin," a manumission document would have coupled the hilye with the erstwhile slave's cins and the declaration that the bearer was now "free like all others of free origin." 60 Here the plaintiffs retained their birth names, but plaintiffs who had been renamed by their holders were often identified by their new names, or by both their birth and slave names even after their freedom had been recognized. The documents thus exhibit the tension between the unfree circumstances in which plaintiffs entered the court and the affirmation of free legal status they had secured when they left.
The litigation is documented in several distinct phases. First, the plaintiffs' initial statement: that twenty days before the kadi heard their suit, two passing janissaries abducted them on the road to the Danube Delta-town of Izmail (in present-day Ukraine) and took them to Istanbul. Next, the defendants are questioned and asked to respond. As the record states, it was already clear from the plaintiffs' language that they were Moldavians-they probably spoke through interpreters-but the defendants' denials, with one claiming that he thought he was buying a Rus' woman, set up the evidentiary phase of the litigation, and the burden of proof rested with the plaintiffs. No one "of free origin" possessed identity documents in this period-only manumitted former slaves had such papers-but even if the plaintiffs had possessed letters of safe conduct or business contracts that might support their claims, written evidence would not be dispositive. Corroborating in-person witness testimony was required to overcome a defendant's denials. 61 In this case, two Moldavian converts to Islam (identifiable by their patronymic) resident in Istanbul testified to the plaintiffs' Moldavian and thus Ottoman origins. Following the investigation (tezkiye) of their bona fides, their testimony was accepted and the judge delivered his verdict.
In most successful freedom suits, testimony came from friends, neighbors, or local officials; ideally the witnesses would hail from the same place as the plaintiff. However, when witnesses from the trafficked person's town or region were 59 Sakk-ı Hızır, SK MS Haci Mahmud Efendi 939, f. 21a-22a. 60 Sobers-Khan, Slaves without Shackles, 235-74. Ottomanized hilyes were a vestigial flourish, an increasingly vague assertion of dominance that, in contrast with the Arabic practice employed through the 1580s, had little identificatory value. The norms observed by Sobers-Khan in 1570s Galata were gradually abandoned with the switch to Turkish, and details like scars and distinguishing marks were rarely described. 61  identified locally and called in to testify by officers of the court, the fact that they might not be personally acquainted with the victim or their neighborhood-despite their testimony following a script that suggested otherwise-necessitated another evidentiary phase. 62 While court records often mention the investigation of witnesses' bona fides, sometimes another judge would be deputized to travel to the district(s) where the witnesses resided to interview locals who knew them. This judge would be accompanied by another Muslim individual so that the two (being the minimum number) could subsequently testify regarding the witnesses' trustworthiness. By the later seventeenth century this method was sometimes employed, albeit rarely, as a way to back-door zimmi testimony against Muslim defendants, since it created an opportunity for the Muslim investigators to independently verify the truth of zimmi witnesses' assertions, then submit their own supporting testimony in court. 63 Although not always described or required, this witness-investigation procedure was necessary in the Moldavians' case since there were hardly any Muslims or Ottoman officials-and thus admissible witnesses-to be found in the autonomous Ottoman tributary states. Although no relationship to the plaintiffs is mentioned, it seems unlikely that these expatriate Moldavian converts personally knew the unlucky honey-selling villagers. Rather the witnesses here may be "experts" of a sort who, being Muslims themselves, could be called upon as necessary to identify and testify on behalf of captive Moldavians. The standardized testimony disguises what may have been a rather complex investigation.
Similar entries appear in the registers of the Rumeli kazasker, who regularly dealt with illegally enslaved Moldavians and Wallachians in the seventeenth century. Many were brought to court by janissaries or other officials, who testified on their behalf. 64 Officials had probably already identified and taken custody of these people, acting on orders to find them; minimally, official help was extended to facilitate release, including procuring witnesses. Officials sent to help did not always have the best intentions, however, and release from captivity did not necessarily mean a return home. Consider the case of Ahmed bin Abdullah, a Moldavian kidnapped in a Tatar raid on his village during the 1672 Ottoman siege of Ladicin Fortress (present-day Ladyzhyn in Ukraine). When the pasha presiding over the siege learned that the Tatars' captives were Ottoman subjects, he ordered their release, but his sergeant Mehmed took Ahmed into his service as a domestic servant. Ahmed, who subsequently converted to Islam, sued for his freedom in October 1679 when Mehmed tried to sell him to someone else, and he won with testimony from officers of the pasha who had ordered his release in the first place. 65 Moldavians and Wallachians, whose consistent identification by cins in the records was necessitated by the unique political status of the autonomous vassal states they inhabited, were one of the rare exceptions to the standard Ottoman scribal practice of deracinating those "of free origin." Their suits serve as a potent reminder that original freedom was explicitly conceived as a function of juridical subjecthood, which the manuals' headings themselves emphasized. Indeed, otherwise enslaveable individuals could acquire Ottoman subjecthood and thus original freedom, as the cases of immigrant converts to Islam, whose cins/asl were likewise noted, reveal.
Those who came of their own volition to Ottoman territory to be "honored by the glory of Islam" were to be warmly welcomed, exchanging "enemy infidel" status for that of the Muslim "of free origin" once they crossed into the Abode of Islam (darülislam). 66 However, border guards and bounty hunters on the frontiers both east and west sometimes failed to recognize, or chose to ignore, the political and confessional transformation that made such people off-limits for enslavement. The central government admonished officials not to detain immigrant converts, but the scenario was sufficiently common for one sixteenth-century sukuk manual to include an example of how to handle an immigrant convert's freedom suit, and they appear in the court records as well. 67 For example, a Mingrelian (from present-day Georgia) named Cafer ibn Abdullah sued in the court of Galata in 1639 after he was arrested under suspicion that he was a fugitive slave. Cafer claimed he had entered darülislam willingly in order to convert and was neither a fugitive nor an enslaveable infidel, but a free Muslim and the court agreed after two high-ranking military officers attested to his prior conversion. 68 The vulnerability of immigrant converts, especially during the times of conflict that often precipitated their journeys, was tangible on the Ottoman side of the frontier as well. Free Ottoman subjects who chose to convert to Islam and relocate, some of whom sought employment with the army, might still be abducted and sold, despite rulings from Ottoman religious-legal authorities explicitly prohibiting the enslavement of traveling zimmis and immigrant converts alike. 69 The upshot was that anyone residing along the frontiers of the empire was vulnerable, regardless of religion or subjecthood, and dislocation and disorder heightened the danger.
As we have already seen from Mehmed bin Abdülcelil's case, long-distance travel exposed Ottoman subjects to peril from hunters of fugitives, who evidently employed wide-ranging ethno-racial profiling to identify targets who appeared "out of place." Many jurisdictions had a dedicated officer (zabıt-ı avabık or yavacı) attached to the court responsible for apprehending fugitives and lost livestock, but anyone could bring a fugitive to the local court for a cash bounty. Those unclaimed after a period, usually three months, would be auctioned, with the proceeds shared among the court's officers. Needless to say, that anyone could arrest fugitives created both a pretense and an excuse for kidnappers caught in the act. Although the immediate resale of suspected fugitives was illegal, neither was it uncommon, as countless ownership disputes involving resold, legally enslaved people attest. To facilitate the trafficking of free people (both "of free origin" and previously manumitted), kidnappers could conspire with unscrupulous court officers, creating a paper trail when the purported fugitive went unclaimed and was subsequently sold. Through the first half of the sixteenth century, the apprehension, interrogation, and reclaiming of fugitives and the payment of bounties to private individuals were routinely recorded in the Istanbul region's court registers, and it is abundantly clear from these sources that bounty hunters violently coerced Ottoman subjects into admitting slave status. 70 Freedom suits often do not record the circumstances by which plaintiffs fell into captivity, but most were probably not previously acquainted with their kidnappers. Yet significant numbers of female plaintiffs knew their traffickers intimately. For instance, Fatima bint Hasan, an inhabitant of the janissary neighborhood in Buda, was engaged to her enslaver, the cavalryman Memişah. In 1592, Memişah brought the unsuspecting Fatima all the way from Hungary to Üsküdar, on the Asian side of today's Istanbul, where to her surprise he left her at the home of a silk weaver named Yusuf and then departed on campaign. After Memişah left, Yusuf sold her for a horse and a few silver coins, and Fatima was taken by boat to Bursa, where she was held "for many days," until Yusuf returned and brought her back to Üsküdar. Both had been summoned to court. Evidently Yusuf's neighbors, who testified that he was "not a good person," had become suspicious and reported him to the authorities. Upon questioning, Yusuf admitted that Fatima was "of free origin," but he claimed that he was just helping Memişah, who had told him that "she should stay with you, and when my sister's husband Mehmed comes, give her to him." Yusuf maintained that he had simply done what had been asked of him, and that when he was ordered to retrieve Fatima, "I went back to Bursa and brought her back, but I didn't sell her." A number of Muslims from the neighborhood contradicted Yusuf's tidy story, however, testifying that he had sold Fatima and that he had often facilitated human trafficking. 71 At this point, the court had pieced together only part of the story, for while it had determined that Yusuf had sold Fatima and that she was "of free origin"-further testimony on this point was unnecessary because Yusuf had not denied it-the nature of her relationship to Memişah was still unclear. About two weeks later, on 30 April 1592, Fatima was back in the Üsküdar court recounting her unfortunate journey to the banks of the Bosphorus. Previously referred to as "Fatima bint Abdullah," her patronymic implying conversion or unknown parentage, she now declared that she was the daughter of Hasan and Ayşe from Buda, that Memişah had taken her for marriage but then on the road to Istanbul had decided to treat her as a slave and resolved to sell her when she resisted. In the intervening weeks, officers had tracked down Memişah and asked him about his erstwhile bride. Now three sword bearers testified that Memişah had informed them that Fatima "is his betrothed, not his cariye (concubine)," while the commander of cavalry recruits and two others testified that Memişah told them that, "if she's my betrothed, I release her, and if she's my cariye, let her be manumitted; give her to someone nice." 72 Memişah's declaration meant that Fatima would not be returned to her enslaver, but the record does not say whether he suffered any consequences, nor where Fatima went next. Despite confirmation of her freedom, she was probably not the arbiter of her own fate. Like Fatima bint Hasan, Emine bint Hüseyin was betrayed by a spouse, but as a resident of Istanbul, she did not have far to travel after winning her freedom in the RSM in June 1649. One and a half years earlier, her husband, Hamdi bin el-Hac Ismail, had dragged her to the market and sold her "as if he were her master" to a slave dealer (esirci), who promptly sold her to a certain Mehmed Efendi. Emine's husband pocketed 180 riyal guruş for his wife, while the esirci sold her for 350 riyal guruş, an enormous sum. In the end, Emine was tracked down by her family, who brought suit on her behalf. On that June day, three elite witnesses from her neighborhood testified while her mother stood by her side. 73 Although successful suits appear in the registers far more than failures-again, plaintiffs won 92 percent of recorded freedom suits-continued bondage may well have been the fate of the majority of those trafficked. As always, witnesses were the key. The case of Şahbaz bint Abdullah reveals how circumstances might enable the enslaved to leave a mark in the registers even while failing to secure freedom. On 6 December 1679, Ismail Bey bin Ahmed, a resident of the Cafer Ağa neighborhood in Istanbul, appeared in the court of Eyüp and stated that Şahbaz had fled twelve days earlier after stealing his watch and other belongings. He subsequently located her, probably with the aid of fugitive hunters, in the home of Şahin and his wife Drako, Christian residents of Eyüp, and now Ismail had come to court to retrieve Şahbaz and his stolen property. Şahbaz responded that she had been "seduced" by the zimmi couple into taking Ismail's possessions and fleeing. She acknowledged that Ismail "has a right to his things," but then she told her story.
Originally from Premedi (Përmet in Albania), Şahbaz declared that she was a Muslim "of free origin" and accused Ismail of holding her illegally. "My husband Mustafa brought me to Edirne and subsequently sold me into slavery," she said, "and I was passed around (tedâvül-i eydî) until I fell into the possession of the aforementioned plaintiff Ismail Bey." Even accounting for the scribe's reformulation of her words, the phrasing is evocative of the humiliation and suffering wrought by repeated transfers-of sufficient quantity and rapidity that neither she nor the scribe saw fit to recount them-and the sexual abuse to which she was and continued to be subjected. But when Ismail denied her claim, she was unable to supply witnesses. Here we might wonder whether the court permitted her to try, for twelve days certainly would not have been enough time for her to send word to family or former neighbors in Albania. But her attempt to escape with the aid of the Christian couple, whom she may have paid to smuggle her out of town, suggests that she had already concluded she was unlikely to secure her freedom through the courts. After Ismail swore that he was unaware of her being of free origin, the court ordered Şahbaz-certainly not her given name-not to raise the issue again without evidence. 74 The door to future litigation was not formally shut if she could locate admissible witnesses, but on that day Şahbaz would have to return to Ismail's home.
Failed attempts to sue for freedom were often transformed in the registers into confessions. For instance, Gülistan bint Abdullah, described as Hungarian, declared in the court of Istanbul in June 1676 that she had previously sued for her freedom 73 RSM 80, f. 63a (7/C/1059). For a comparable case, see Tamdoğan, "La fille du meunier." 74 Eyüb 90, f. 31b (3/ZA/1090). while she was the legal property of Fatima bint Süleyman, a slave dealer from Belgrade, when Fatima sold her to el-Hac Murad bin Ibrahim. At the time, Gülistan had declared that she was from a small village outside Belgrade and that the esirci Fatima knew this when she contracted the sale to Murad. But now Gülistan "canceled" (mubtılayım) her earlier suit, and she acknowledged (ikrar) in court that she was the legal slave of el-Hac Murad, "may he use [me] however he wills." 75 Similarly, in June 1686, Ayşe bint Abdullah recalled that she had sued for her freedom when she was sold three months earlier to Hüseyin Beşe, but now in the Istanbul Bab court she admitted that she was "a slave of Rus'ian origin" and canceled her previous suit. 76 Had Gülistan and Ayşe lied when they first sued in order to short-circuit their sales? Were they coerced into withdrawing their suits, or were they ultimately convinced, by the judge or their presumptive owners, that they would fail in their efforts? Captives undoubtedly made desperate attempts to sue for freedom even if they were not, by Ottoman legal standards, "of free origin"-if nothing else suing might suffice to scare off a particularly undesirable buyer-but Gülistan and Ayşe's cases may represent the experiences of many illegally enslaved Ottomans who came to court and were unable to substantiate their claims. By making these women appear in court to publicly acknowledge their enslavement-seemingly standard practice by the late seventeenth century-their holders protected their reputations and inoculated themselves against future litigation and financial claims if the women ever sued again. 77 After all, the records are full of instances of buyers suing sellers for their money after successful freedom suits, and the sukuk manuals provide many sample entries of this litigation. 78 One manual includes an "example of the return [of price] at the third level [of remove from the initial sale]," which gives some indication of how frequently some slaves, especially women, were sold, and the long, complicated chains of litigation their successful freedom suits could trigger. 79 Repeated sales made establishing provenance increasingly difficult, especially when perpetually peripatetic Ottoman male elites were involved. 80 When the trail finally ran cold, a former owner unable to find the previous seller could be left holding the bill, years after he had bought and sold someone.
That is, unless that person had been coerced into denying their freedom and declaring that they were a slave, as Mehmed bin Abdülcelil had to Allahverdi, or the buyer had two reputable associates who were willing to lie to that effect in court. Several sukuk contain, in addition to examples of freedom suits and litigations to recover purchase prices, entries like "example where it is proved that he said 'I'm a slave, buy me.'" 81 Consider the example of Mehmed bin Yusuf, an adolescent renamed Piyale, in the seventeenth-century Sakk-ı Fındıkzade. Abdülkerim encountered the captive "Piyale" in the market in Arnavud Belgrad (Berat in Albania) and interrogated his potential purchase: "Is it your wish that I buy you? Will you legally dispute my ownership with [a suit] 'of free origin' or with the appearance of another who has a claim to you? Do not deceive me, are you truly a wholly-owned slave?" In response, Piyale supposedly admitted that he was an enslaved Croat, and Abdülkerim bought him from his captor, Mehmed Beşe. Soon, Abdülkerim sold Piyale to Ali Ağa, who sold him to Mehmed Ağa, and while in Mehmed Ağa's custody, Piyale successfully sued for his freedom and returned to his old name, Mehmed bin Yusuf. 82 Mehmed/Piyale may have been eight or nine years old when he was bought in Berat. Probably still uncircumcised when he was abducted, he certainly cannot have been expected to understand the weight of his words in the market. But when Mehmed bin Yusuf secured his freedom, he triggered a chain of litigation that ultimately boomeranged: Mehmed Ağa sued Ali Ağa, and Ali Ağa sued Abdülkerim, and finally, when Abdülkerim could not find Mehmed Beşe, the man who sold him "Piyale," he sued the erstwhile Piyale himself, Mehmed bin Yusuf. Three years had passed since their initial meeting. Once in court, Mehmed bin Yusuf did not deny that he had said he was a slave, but he claimed he did so under duress, that his captor had beaten and kicked him and threatened to do worse if he said otherwise. He had been a terrified child, but it did not matter. Unless he could summon his own witnesses who could testify that he had been forced by Mehmed Beşe, he would have to compensate Abdülkerim. 83 Albeit reformulated into the court's idiom, these records reveal that brief interviews in the market were common, and prudent buyers inquired about the origins and legal status of captives to inoculate themselves against future claims. If persons of free origin had the wherewithal to protest that they were Ottoman subjects, they would likely be disbelieved-though this may have been how many managed to bring their cases to the attention of the authorities-and they risked retaliatory violence from their captors. If, on the other hand, they gave the affirmative answer that seller and buyer expected, they might temporarily improve their situation but expose themselves to future ruin if they ever won recognition of their original freedom. Proving original freedom did not automatically result in freedom from servitude; it might simply swap the social status of slavery for the ignominy of debt bondage. 84 The blurred boundary between free and enslaved is further exemplified by the numerous domestics "of free origin" forced to sue when their employers attempted to sell them. 85 In contrast with the stories narrated above, no violent kidnapping was necessary, just a wealthy household's exploitation of vulnerable, socially marginal retainers. For these plaintiffs who had only recognized themselves as illegally enslaved at the moment of their sale, would winning "freedom" in court mean unemployment and homelessness? If so, the ultimate aim of litigation may have been to renegotiate their position within the household rather than to escape it. Such cases underscore the need to look beyond the free/enslaved binary, beyond legal status to the social practice of slavery.

Conclusion
The distinction between free and enslaveable in the premodern Ottoman Empire was politico-legal, not ethno-racial; the phrase "of free origin" took the form of a cins label but subordinated any other that might apply. Even in bondage, plaintiffs contended, and their witnesses testified, "They were not and had never been slaves," and time after time courts agreed. Whereas Ottoman slave-buyers undoubtedly had cins preferences, there was not, legally speaking, any cins that was de jure enslaveable. Changes in political or religious status-that is, conquest or conversiontransformed individuals or whole communities from "enemy infidels" into Ottoman subjects "of free origin," and the reverse.
While 27 percent of suits (twenty-one) in the sample originated in the east-and brigands frequently kept women and children in thrall in Anatolia during the chaotic seventeenth century 86 -the fact that the majority came from the European half of the empire permits three key conclusions. First, the trade in illegally enslaved Ottoman subjects worked best when camouflaged by the larger traffic in licitly acquired captives. Besides the Caucasian borderlands, in the seventeenth century this was most practical in the European provinces, where the difference between the protected and the enslaveable was purely juridical, with the faith, language, and cins often identical. Second, many were abducted under the cover of military movements on land or sea from areas on the way to, or not far from, the frontiers, or which hosted sizable garrisons; young women in fortress communities were at significant risk. Third, the proportion and distribution of plaintiffs hailing from Europe says more about which regions' captives were able to secure witnesses, and where the central government was capable of projecting authority, than it does about the actual geographical distribution of illegal slaving. Traffickers' efforts to conceal their activities worked, obscuring their full scale.
Over two-thirds of litigants were women, most employed in domestic and sexual service. Unsurprisingly, in 86 percent of cases (sixty-eight) the defendants were men, most of them economic, religious, and military-administrative elites, and all but three were Muslim. All but one of the female owner-defendants in the sample faced off against women. Not only were female captives preferred by buyers, but they were also easier for traffickers to pass off as legal slaves. After all, the only things that ostensibly distinguished Muslim and non-Muslim women in the Ottoman Empire were their names and clothes, both of which could be changed by slavers and buyers. Scholars have previously noted the similarities between slavery and marriage in Islamic jurisprudence, but the sources reveal how easy it was for husbands to convert their wives into chattel once removed from the protective embrace of their home communities. 87 The power of Ottoman subjecthood was no match for the authority of a husband over his wife.
The constant churn of the market, driven largely by male demand, incentivized illegal slaving, but it also enabled some captives to bring their plight to the authorities' 86 See Leslie Peirce, "Abduction with (Dis)Honor: Sovereigns, Brigands, and Heroes in the Ottoman World," Journal of Early Modern History 15, 4 (2011): 311-29. 87 Zilfi, Women and Slavery; Kecia Ali, Marriage and Slavery in Early Islam (Cambridge: Harvard University Press, 2010). attention. Because major purchases were often witnessed and registered with the courts, the moment of sale could provide the opportunity to sue for freedom. For some, that opportunity arrived within days. For many, it took years. But sale was not the only path to court. Some captives were brought in by bailiffs or janissaries tasked with finding them. Others were tracked down by their families, and though parents did not testify, their presence in the courtroom was often noted. 88 Some escaped and declared their freedom after they were caught, like Mehmed bin Abdülcelil at the beginning of this article. Some were actually sued first by their holders for disobedience, and, once in court, turned the tables by declaring their freedom. 89 How were witnesses summoned and court fees managed? How often did bribery and witness tampering derail (or facilitate) freedom suits? Above all, what happened next for plaintiffs? Many questions remain. Even though we can locate the agency of the enslaved in these cases and discern the shape of their legal arguments, we can recover only faint echoes of their voices. As for those turned away by the courts, the sources are wholly silent. The standardization, shaped by sukuk, of Ottoman legal documentation beginning in the late sixteenth century reveals these cases' prevalence and, simultaneously, conceals their true scale.
The illegal enslavement of Ottoman subjects undermined the sultan's promise to provide protection and justice, for which reason officials sometimes went to great lengths to find and free captives and periodically tightened control over and surveillance of slave dealers and brokers. Nevertheless, we must conclude from the inclusion of sample freedom suits in nearly every legal praxis manual that illegal slaving was not only unremarkably commonplace but tacitly accepted as an unavoidable consequence of maintaining a robust slave market, which was deemed essential for the maintenance and reproduction of elite Ottoman households. The basic fact, exploited by slavers and tolerated by officials, was that proving "original freedom" via Ottoman subjecthood was functionally impossible for many if not most inhabitants of Ottoman domains. For every plaintiff who won her freedom (if not necessarily her liberty), more may have been sent back to their captors. The Catholic corsairs and Cossack raiders who menaced the empire's frontiers in the seventeenth century were the most visible threat to Ottoman subjects' freedom, but the internal danger was more widespread and insidious. Scholars of slavery, long reliant on legal sources and primed to explore the subject through the legal lens, must contend with the fact that the subjecthood of captives and the legality of their enslavement mattered a great deal-except when it did not matter at all.