Halil Inalcik. "Servile Labor in the Ottoman Empire"

in A. Ascher, B. K. Kiraly, and T. Halasi-Kun (eds), The Mutual Effects of the Islamic and Judeo-Christian Worlds: The East European Pattern, Brooklyn College, 1979, pp. 25-43.

The long prosperity of the slave trade in Islamic lands can be attributed to conditions peculiar to Islamic culture. First of all in medieval Islam from the time of the Abbasid caliphs on slaves were employed in large numbers in the militias of Muslim rulers. Slaves were also used as labor force in the urban crafts and on the big estates belonging to the state or to large land owners. Furthermore, following the example of the palace, the upper class and even the well-to-do among the non-Muslims, had extended households with large numbers of domestic slaves.

In the Service of the State and Military Class

The ghulam or kul system, in which slaves were trained as loyal servants to be employed primarily in implementing the central power of the sultan as agents and soldiery, had an unprecedented expansion under the typically military centralist sultanates of the Mamluks of Egypt and the Ottomans in the period of the thirteenth-sixteenth centuries. With the rise of their centralist state, the Ottoman sultans were faced with a growing need for men for their kapi-kulu, palace servants and divisions of the standing army at the Porte, and, since war did not bring slaves in sufficient numbers, they had to resort to the unusual method of the devshirme, a levy of boys from among their own Christian subjects, the reaya'. In general the Ottomans did not regard the devshirme as enslavement, but rather as one of the extraordinary services imposed by the state in an emergency. The levied boys, attached to the Janissary corps, were first employed as labor in a number of public works ( construction work, transport works etc. ) in the capital and in Gallipoli before they became Janissaries. In the classical period between 1300-1600, the Sultan's kapi-kulu, recruited from the devshirme boys as well as from among prisoners of war and slaves bought for the Sultan, increased in number considerably: 15-20 thousand under Mehmed II (1451-1481), 60 thousand in 1568 and about 100 thousand in 1609. The absolute power of the ruler, we are told in contemporary Ottoman sources, rested upon his having slaves in his service in the army and the administration.

That the members of the Ottoman ruling class, the Sultan in the first place, took special care to increase their slave retinue can be related to the fact that, in the frontier society of the early Ottoman state power rested on the ability to muster as many ghazi fighters as possible from tribal companions, adventurers from outside, or slaves. It was true the use of warrior slaves was not confined to the state. The grandees also tried to maintain households and retinues as large as possible since this meant, through their patronage rights, wala, an extension of their influence and power, in-as-much as many of their slaves were destined to get important offices in the Empire later on. These patrimonial factions were not infrequently the real actors in political feuds in the Ottoman Empire. In Istanbul, capital of the Empire, at least one-fifth of the population was estimated to be kuls of the Sultan and of the grandees. In other cities kuls, though in smaller proportions, made an important part of the urban population.

The demand for warrior slaves in Ottoman society was intensified by the fact that members of the ttmar holding army in the provinces, from the beglerbegi, governor general, down to the simple sipahi, cavalryman, were required by regulation to train and to bring to the Sultan's army a certain number of cavalrymen in proportion to the amount of their timar revenue (for a sipahi, one for each 2 or 3 thousand akce, for begs, one for each 5,000 akce). These auxiliaries were in their origins mostly slaves captured during the raids in Christian lands. A beglerbegi, governor general, was required by regulation to maintain a military retinue of at least two hundred men in addition to his household slaves employed in domestic services. The begs strove to increase the number of their retinue beyond the regulation since that brought special favor from the Sultan.

However, the ruling class did not own as many domestic slaves as military. On the basis of a list of estates belonging to members of the ruling class kept in Edirne between 1545 and 1659, the following data was collected: out of 93 estates, 41 had slaves.

The total number of slaves in the estates was 140, 54 female and 86 male. 134 of them bore Muslim names, 5 were not defined, and 1 was a Christian woman. Some of these slaves appear to have been employed on farms. In conclusion, the ruling class, because of extensive use of warrior slaves and because of its own high purchasing capacity, was undoubtedly the single major group keeping the slave market alive in Ottoman empire.

In the Crafts

Concerning the slave labor employed by the crafts in the cities we have valuable evidence from the records of the estates of the deceased kept by the judges, kadis, in the Ottoman cities. The following observations are based on such records of 721 estates from the second half of the fifteenth century in Bursa.

It was an exceptional case for the rich not to have slaves either in domestic occupations or employed as labor in certain crafts. (The rich composed 15.9 of the cases studied.) Slaves formed the third most important component of the estates, in value, after cash and properties. In the estates of the silk weavers slaves frequently represented the most important part since slave labor was extensively employed in the weaving of gold brocades, velvets, or fine cottons in Bursa.

Slave labor was organized under the system of limited service contract known as mukataba in Islamic jurisprudence. Here are two examples of such a contract:

"In our presence Mahmud b. Seyyidi Ahmed, weaver of taffeta, asserted that he agreed to emancipate his slave Iskender, of Circassian origin (with the described features) upon the completion of one hundred pieces of taffeta equivalent in value to ten thousand akces, and the said slave accepted the contract."

"Khwadje Sinan had previously agreed to emancipate his slave Shirmerd son of 'Abdallah (a convert), of Slavonian origin, upon the completion of the weaving for him of ten brocades known as kemkha-i glistani. Now that he has completed the work he has become free.

Here is a case of a slave weaver who was emancipated and rewarded by his master upon his death:

"Yusuf B. 'Abdallah (a convert) previously slave of Al-Hadj Tannvermish, asserted in our presence one day before his death to the effect that he emancipated his slave Ayas b. 'Abdallah (a convert) of Russian origin, weaver of velvet; and declared in his will that Ayas be given in his possession the loom of velvet with silk and other pertinent things."

This kind of mukataba meant actually to allow the slave to exercise certain rights such as to work independently and to own his earnings so that he would be capable of ransoming himself. Another kind allowed emancipation upon work for a certain period of time without specifying the work. Example:

"Mawlana Seyyid Mehmed of Konya asserted in our presence that he agreed to set free Lutfi b. 'Abdallah (a convert) of Bosnia ( of the described features ) upon service for him for four consecutive years; and the slave accepted the terms

. Mukataba was widely practiced in the Ottoman Empire, as demonstrated by the kadi records. It is recommended by the Coran. It consisted in the master's granting his slave his freedom in return for the payment of mutually agreed upon sums of money. According to some legists it was ransom by the slave of his own person. As an interesting historical example, mention can be made of Mehmed the Conqueror's allowing the Greek prisoners of war to work at the repair of the walls of Istanbul to ransom their freedom. The person subject to mukataba is set free only when his payments are completed. Toward the end a rebate was normally accorded.

Mukataba was a contract binding both sides so that the owner could not change its conditions at the expense of the slave. Since emancipation was considered a charitable act, the owner might make modifications favorable to the slave, such as to shorten the period of service or to give up the work due. Of course at the same time the owner derived certain advantages from mukataba. It guaranteed good and profitable service for a certain period of time since, as a rule, lifetime slaves tended to run away or to be indolent. It was particularly profitable in the silk industry as this required continuously careful expert work especially in brocade and velvet weaving. Wage laborers were not really suited for this kind of work, which demanded a long period of time on the loom for the production of a single piece.

Noteworthy also is the fact that the silk industry in Bursa had developed to such a point that it exported its costly gold brocades and velvets, not only to meet the growing demand of the upper class in the empire, but also to meet orders from Italy, Poland, Russia and other European countries. There were silk weavers in Bursa having forty or more looms at one time who can rightly be considered as capitalist entrepreneurs, organizing a domestic silk industry for export with slave labor: Al-Hadj Ahmed with five looms and fifteen slaves, Hadjdjl Badr al-Din Ishak with seven looms and eight female slaves, al-Khadje Sinan with six looms and twelve male and three female slaves, were all active in the middle of the fifteenth century. Twelve of the slaves of Al-Hadj Ahmed had a value estimated at 36,000 akces, median price for a slave being 2,000 akces or 50 gold ducats, a rate equal to or below contemporary average prices of slaves in Italy or Egypt. Since slave labor was not cheap it was only in the crafts making high priced luxury goods in great quantity that slaves were employed. Our Bursa documents refer to no slaves in other crafts. In the weaving of cheap cotton goods, another Ottoman export item, peasant labor in the coun- tryside and widows and children in the towns were used through- out Anatolia as the cheapest labor available.

Finally, hiring out slaves was legal. H. Dernschwam, a German visitor to Turkey in 1555, relates the widespread practice in Istanbul of hiring out slaves: Many people made a livelihood out of hiring out their slaves for 7 or 12 akces a day as day laborers (then 60 akces equalled one gold ducat; a slave's daily expense was estimated at 1 1/2 or 2 akces ).

In Commerce

Bursa documents of the fifteenth century also tell us how the freed slaves, 'atik or ma'tuk, occupied an unusually important place in the economic life of Bursa as rich silk manufacturers and merchants engaged in distant caravan trade, in money exchange, in usury and in tax-farming. In that city in 1477, 61 out of 402 persons whose estates were recorded in the kadi registers after death ( 34 male and 27 female ) were freed slaves.

Slaves and freed slaves were often employed as commercial agents by merchants in distant trade ventures. Special guarantees under the stipulations of Islamic law of toal, patronage rights of the former master, must be emphasized in this connection. The following is an interesting instance: In 1480 Balikcizade of Bursa and Khvadje Mehmed, freed slave of Khadjadj Koci, made a partnership investing equal shares with capital in the amount of 545,000 akces ( approximately 60,000 gold ducats) for the purpose of an import-export trade with Egypt and Syria. The operation was conducted mainly through their slaves who made several trips via Antalya ( Satalia) between Bursa and Egyptian and Syrian cities. In his will Balikcizade emancipated upon his death three of his slaves whom he had used as commercial agents (in addition one eunuch and one female slave were to be emancipated with grants of money, while three female slaves with children, umm al-walad, from him were to be freed automatically at his death. Of course, as in the case of Balihkcizade most, if not all, of these slaves were employed in domestic occupations. The peculiar stipulations of Islamic law gave rise to a paternalistic type of masterslave relationship which fostered strong social ties especially where domestic slaves were involved.

In the Agricultural Sector

Here slave labor became predominant on the big estates, which were in the form of ciftliks belonging to the state, the wealthy, or powerful members of the ruling class, or in the form of awkaf, pious endowments. The main reason for this was that the re'aya, free peasants registered in the state survey books for taxation in a defined area, could not, under the law, be used in the newly established farm lands. The majority of the privately owned ciftliks and many of the state farms and trusts were formed on the uncultivated waste-lands, pastures, and uninhabited lands with servile labor. Furthermore, such ciftliks were market oriented, set up essentially for the purpose of profit. Cattle and crops were sold to nearby peasants in need or shipped to the ports for export or for provisionment of the Ottoman cities. As in the Western plantation system, maximum rentability was the main concern of the ciftlik owners or trusts; slave labor, free of any of the restrictions to be observed under the law for the re'aya was found to be the most suitable. Slaves besides were comparatively cheap and readily available during the period of Ottoman expansion. It should also be pointed out that organization of production on the ciftliks adjusted itself to labor conditions: stock raising became predominant on most of the ciftliks as it required a minimum amount of labor for the highest rate of profit. Ortakdjlik/sharecropping with equal shares between the owner and slave was a general practice on the ciftliks while if cultivated by a peasant of the re'aya-status, the owner's share could not exceed as a rule one-eighth of the produce. Lastly slaves constituted a capital investment which was easily, and most of the time profitably, convertible into cash. It should be noted that in all this members of the Ottoman ruling class were following practices long established in the Islamic world.

Under the first Ottoman Sultans servile labor appears to have been employed to form extensive cattle and sheep ranches in Bythinia. Big farms thus formed by the grandees were mostly turned into pious endowments. As an inalienable part of the farms slaves were entered in the endowment documents often by their names. In transactions they are treated in the same way as other properties. In some endowment deeds it is made clear that slaves settled in the farm or village were prisoners of war captured by the founder. Over time some of the farms thus formed developed into villages and the descendants of the slaves in them were always separately registered as sharecropping slaves (ortakdji kul). In some other cases it is reported that slaves ran away and farms or villages were left in ruins.

The best example of the state's use of slave labor in reclaiming an abandoned agricultural area is Mehmed the Conqueror's attempt at settling with slaves a large area called Khasslar encompassing 163 villages around Istanbul and Pera between 1453 and 1480. Actually it was a part of his larger plan for the reconstruction of Istanbul. By placing under state ownership the arable lands under cultivation in this area he intended to contribute to the provisioning of the city, to create new sources of revenue for the treasury, and to keep the neighborhood of the capital in order and safety, as is pointed out in the regulation made in 1498. Mass deportation of enslaved peasants from enemy territories was used for this purpose. This was not only because the Conqueror planned a rapid recovery of the area but also because settlement of reaya deportees, surgun, from his own territories proved to be difflcult to carry out and ruinous for the areas from which the deportation was made. It was impossible to go too far in this operation because it violated the basic rights of the re'aya, Muslim and non-Muslim, by forcing them to stay in their new settlements. It should be added that shortage of agricultural labor was a general phenomenon in this period.

Actually, in the Khasslar, on land that belonged to the imperial treasury, the enslaved Greek population, as well as slaves from newly conquered Bosnia, Serbia, and the Morea, made up the majority of the population, along with the deportees, surgun of re'aya origin, sedentary or nomadic, and the group of the ordinary re'aya. In 110 villages out of 163, slaves were in the majority. In 1498 there were altogether 1974 slaves left in these villages.

Since it was the prime concern of the administration to maintain the productivity of farm units, changes leading to the freedom of slaves (even in cases permissible in the religious law) were not tolerated in the Khasslar. There were however, many examples of slaves who became freedmen by illegal means, the principal one being the bribing of the emin, or official in charge of administering the Khasslar. The money paid to emancipate a slave, aghirlik, appears to have originally been a compensation defined by the regulations but diverted by the emin to his own pocket.

In the survey book of 1498 there were many who claimed to have been freed but who could not establish their emancipation by any evidence acceptable to the surveyor-inspector. As an occupation other than agricultural work actually made easier a change from slave status, in 1498 slaves were forbidden to engage in jobs such as fishing or carrying wood or lime in carts to the city, profitable occupations in the neighborhood of Istanbul.

There were greater opportunities for the female slave than for the male slave to change her status. Marriages with the re'aya, free peasants who paid an aghirlik to the agent to obtain his consent, frequently occurred. In 1498 the government inspector observed that this widespread practice and the unwillingness of the female slaves to marry the khass slaves, inhibited the continuation of slave families and the proper cultivation of the farm units in the Khasslar. Under religious law the children of a free man were to be free. Even if male slaves married free women their children were to be free. Apart from the decrease, in the long run, of the slave population there was also a disruption of the agricultural work, which was based on family-labor on farm units. These points were stressed by the inspector in his report to the Sultan. In the new regulation of 1498 marriages of slaves with free men were forbidden and forced marriages between the slaves, which were provided for under the religious law, were enforced. Exceptions were made for the widow whose son was able to maintain the agricultural work and for anyone who had an acceptable excuse for not wanting to get married. License for marriage with a free man was given only when there were more female slaves than needed, and when the free man agreed to undertake to continue the sharecropper's work on the land. For such marriages a specific license of the Sultan, sureties, and payment of a marriage due called aghirlik were demanded.

Comparable in its basic features to the Western colonate system, the use of war captives in agriculture, however, had a limited application in the Ottoman Empire. Under Suleyman I (1520- 1566) in the central part of Rumeli encompassing Thrace and Macedonia, the slave agriculturists numbered only 6021 men, including those in the Khasslar, about 2 percent of the whole population of the region; and, in the province of Anatolia 1981 men. The number must have been somewhat larger in the previous period since the state had to change slave status of some communities into that of re'aya -- a change which seemed to take place as a result of the great difficulty of keeping under control these slave colonies, which constantly dwindled away through marriage and intermingling among the overwhelming re'aya majority around. In fact, in the subsequent surveys in the sixteenth century, the slaves in the Khasslar are referred to merely as sharecroppers. Over time slave agriculturists everywhere in the empire were to be identified with the re'aya masses and disappear.

Slavery, it can be safely said in conclusion, was an institution of vital significance for Ottoman society. Not only the state organization but also various segments of the economy -- the silk industry, ciftlik agriculture, distant trade, as well as the extended household-type family of the upper class -- all rested upon slavery. It must be emphasized however that all were dependent on a regular large-scale supply of slaves from outside, since slavery in a Muslim society could maintain itself only with importation.

Islamic jurisprudence recognized only one category of slaves -- those born in slavery or captured in war. Those Muslims born in slavery or converted while slaves could remain in slavery, but no free Muslim or dhimmi, non-Muslim subject of the Sultan could be reduced to slavery. When a slave woman had a child by her master the child was born free. Islam also encouraged emancipation as a work of piety. The children of a female slave are free on her emancipation. A grant of enfranchisement taking effect at the master's death, a practice called tadbir, favored by religious law and widely used as the Ottoman kadi records and testaments demonstrate, was perhaps the most important factor in eroding the slave population. Also contractual enfranchisement as seen above was widely practiced in the Empire.

Supply of Slaves

In the period extending from about 1260 to 1390, the age of the great expansion of the Anatolian Turks into Bvzantine territories in western Anatolia and Thrace and into Macedonia and Bulgaria, captives flowed into the Ottoman lands in great numbers. It can even be added that during this period the great demand for slaves from the Islamic hinterland and rising prices played an important part in the extension of the raids and consequently the rise of the prosperous ghazi principalities in Western Anatolia. Paradoxically enough not only for the slave markets in Italy but also for agricultural labor in their Levantine colonies, Venetians and Genoese too became regular customers of these principalities. As a result of its growing need for slaves, especially during the periods of rapid growth such as experienced under Bayezid I (1389-1402) and Mehmed II (1451-1481), as well as when internal crises slowed down the flow of slaves, the Ottoman state experienced shortages in slaves, and as explained above, had to resort to extraordinary measures, such as the levying of boys from the non-muslim reaya, and the conducting of mass enslavements and deportations from the newly conquered lands.

On the other hand the rise of the Ottoman Empire affected the international conditions of slave trade in the Mediterranean in general. The slave traffic to Italy from the Balkans through Dubrovnik stopped. Due to the strict control of the traffic through the straits after the Ottoman conquests of Istanbul (1453) and Caffa ( 1475) and the prohibition of traffic in Muslim slaves, mostly Turco-Tatars of the steppes north of the.Black Sea, the slave trade betveen the Crimea and Egypt slacked off or changed its early character. As is well known, this trade, first in importance in the commerce of the Black Sea, was a Genoese monopoly. The new circumstances resulted in the ruin of the Genoese prosperity in the Black Sea, caused higher prices for slaves throughout the Mediterranean, and the replacement of Turco-Tatar slaves with Caucasian and Russian slaves.

The impact on the Mamluk system has not yet been the subject of a proper study. What is certain is that the Mamluks now offered unusual prices for Caucasian slaves which greatly encouraged the trade in Circassians and other Caucasian peoples. On the other hand, the new situation caused the prices to treble in Italy, and seems to have reinforced, among other factors, the anti-slavery feeling in Western Christendom, and undermined slavery as an institution in sixteenth-century European society. Now the only exception in Europe was the great need for galley slaves, who were provided mostly by European corsairs from the territories under Ottoman rule.

In the fifteenth and sixteenth centuries, before the Crimean Tatars became the main suppliers of slaves, the Ottoman ghazis or akindjis, raiders, on the frontier areas in the Balkans and Central Europe met the huge demand for slaves in the Ottoman market. Suleyman's reign ( 1520-1566 ), the zenith of Ottoman power, witnessed also the great extension of raiding and enslavement activities by the Ottomans. Capture and sale of slaves usually brought to an individual akindji or Ottoman soldier a sizeable income -- a strong incentive for him to join the raids or campaigns. It also furnished an important source of revenue for the Sultan's treasury, since, as we know, one-fifth, pendjik, of the captives or of their market value belonged to the Sultan. The captives always brought a good price in the cities' slave markets, which were organized originally and constantly supervised by the state.

An eye witness, Konstantin Mihailovic, gives the following description of raids organized in a frontier center in the fifteenth century:

"The Turkish raiders are voluntary -- of their own will they ride on expeditions for their livelihood.... They live by means of livestock and raise horses.... If any of them does not want to go on a foray himself, he will lend his horses to others for half (of the spoils); if they win some booty they accept it as good, but if they bring nothing, then they say 'We have no gain, but we have great works of piety, like those who toil with us and ride against the Christians, because we support one another.' And whatever they seize or capture, whether male or female except for boys, they will sell them all for money. The emperor himself will pay for the boys."

The slave merchants, esirdjis, working in the frontier centers or following the Ottoman armies, bought captives quite cheaply and brought them to inland markets, the most important of which were in Iskub (Skoplje), Edirne, Bursa, and Istanbul. By the end of the fifteenth century the Bursa slave market appears to have been the liveliest as demonstrated by the kadi records. Even the Sultan sent slaves to be sold at Bursa where, obviously the best prices were expected. Persian silk merchants were among the best customers. Apparently Bursa replaced Sivas in the slave trade in the Middle East during the Ottoman period. Edirne appears to be the main slave market in the Balkans. B. de La Brocquiere on his journey in Rumili in 1432 saw such a train of 25 captives led by slave merchants to Edirne. Later in the peace treaty with the Ottoman state Austria made the Porte agree not to allow the slave merchants to roam about the borders and buy slaves captured in violation of agreements. Not only European visitors who met trains of slaves on their way to the Ottoman capital, but also Ottoman historians and epic literature furnish a vivid picture of this activity. It should be noted that the Holy War, and distribution of booty, ghanima, played a major part in early Islamic history, and were regulated in every detail by Islamic law. The Ottomans followed these regulations closely, as they did those on slavery in general.

Despite the fact that the Ottomans themselves used most of these slaves, there is documentary evidence that captives taken by Turkmens of the principalities of western Anatolia and by the Ottomans in the period fourteenth-sixteenth centuries became subject of an export trade to Egypt and Italy. In Venice's Levantine possessions, for instance on Crete, slaves of Balkan origin were employed in agriculture, and in Italy (Florence, Milan) in certain specialized crafts. Italian notarial documents demonstrate that slaves from the Balkans, Greeks, Wallachians, Albanians, and Serbians, appeared in Italy in the fifteenth century (with the exception of the Bosnian slaves who appeared earlier). The same observation is made in the Mamluk kingdom in Egypt. An Ottoman customs register in Antalya (Satalia), on the southern coast of Anatolia, dated 1560, also tells us that while white slaves were still then exported to Egypt and Syria in quite substantial numbers, in return black slaves constituted an important part of the imports from those countries.

It should be noted that particular ethnic groups among the slaves in the Ottoman empire or among those purchased by foreigners became dominant at given periods of time depending on where Ottoman raiding was then intense. In the fourteenth century, Greeks and Bulgarians; in the fifteenth century, Serbs, Albanians, Wallachians, Bosnians; and in the sixteenth century, Hungarians, Germans, Italians, Spaniards and Georgians.

In the second half of the sixteenth century none of these sources of supply could compete with the Black Sea, the importance of which grew in proportion to the decline of the Ottoman akindji organization in the Central European frontiers mainly as a result of the stiffening of Austrian resistance. Now the Crimean Tatars supplied the larger part of the Ottoman market, specializing, so to speak, in the business. Raids and expeditions into Poland, Circassia, and Russia became a regular occupation of vital economic importance for the tribal aristocracy of the khanate, so much so that often their relations with their khans or with the Ottoman Sultans were determined by their policies on this fundamental issue. Muscovite lands became the main field of operations from the time of Muhammed Girey I (1514-1523); and large scale expeditions became regular after 1534 when the legacy of the Golden Horde in the Volga basin was the subject of a long struggle between Muscovy and the Crimea. The Crimeans naturally considered their actions as a Holy War against an enemy who occupied the sister Muslim khanates in the Volga basin and threatened the Crimea itself.

The slave trade was indeed the foundation of the Crimean economy,. Essentially it was the economic pressures resulting from immigration or from drought and famine, a frequent occurrence in thc region, that thrust thousands of men, tribal warriors as well as commoners, into Russia and Poland for raids. Revenue accruing from the sale of slaves, contemporary observers asserted, constituted a real relief for the country at such times. Nogays, pure nomads of the steppes outside the Crimea, were absolutely dependent on the slave trade, selling their captives wholesale to the merchants coming to their headquarters. Most of the slaves were exported to the Ottoman market, but an important part of the captives was employed by the Crimean tribal aristocracy itself as agricultural slave labor to grow cereals for the city of Istanbul. Obviously that was the reason why they sometimes tried to capture and move whole families in their raids. If Evliya Celebi's statement can be trusted, there were 400,000 slaves in the Crimea as against 187,000 Muslims (100,000 of the latter were, he added, commoners and 87,000 military) in 1667.

Raids by the Crimeans into Russia or Poland, usually in small parties from 200 to 1000, were continual occurrences irrespective of the formal peace between the Crimean Khanate and those countries. During their raids, the Crimeans avoided as much as possible the line of defenses which consisted of a series of the fortified towns with garrisons on the frontier. Scattered along an extensive frontier line, the Russian forces were often powerless before large scale Tatar invasions carefully organized under the chiefs of the tribal aristocracy or members of the Girey dynasty. In these raids slaves were considered as real booty, and their safe transport was always the chief concern of the raiders: they usually tried to avoid fighting, and doing anything which might lessen the value of their human chattel.

As under the Genoese, Caffa (Turkish Kefe) and Kerch were the principal trade centers for slaves in the Crimea. The other less important centers in the region were Azov (Azak), Taman, Copa and Sokhum where Tatar, Circassian and Ottoman slave merchants met. Slave trade and taxation in the Crimea and Azov were regulated under special laws which apparently were made on the Genoese models. The slave tax, taken as a rule at Caffa, was quite high: 210 akce, per slave, and with some additional dues reaching 255 akce (about four gold ducats at the end of the fifteenth century) .

At the height of the Crimean raids into Russia and Poland between 1514-1654 captives are reported in unusually great numbers. In 1578 annual customs revenue from the slave tax at Caffa was estimated at 4,463,196 akce. Divided by 255, the highest tax rate, the figure 17,502 slaves per year is obtained. In 1614 following the large-scale expedition of Djanibek Girey Khan each Tatar soldier came, according to Russian sources, with five to ten slaves and prices at Caffa went down as low as 10 or 20 gold ducats per adult male slave, while average price was over forty gold ducats in Edirne during this. About thirty major Tatar raids were recorded into Muscovite territories between 1558-1596.

Whenever a peace with Muscovy, unpopular though it might be, was imposed upon the tribal aristocracy they often forced the khan to lead them in expeditions into Circassia, the second important region for slave raids during this period. Excuses were easy to find since the majority of Circassians were pagans at this time, and their chieftains refractory to the Crimean suzerainty. Sahib Girey Khan (1532-1551) who intensified the anti-Muscovite policy of the Khanate was also responsible for large scale expeditions into Circassia which resulted in mass enslavements. In the 1530s Kansavuk, Circassian chieftain of the Jana agreed to send a yearly tribute of one thousand slaves to the Ottoman Sultan and five hundred to the Khan. Upon his failure to keep his promise, Sahib Girey led an expedition against him in 1539, and took, according to his court chronicler, 50,000 captives. We learn from the same source that in his subsequent expedition against Kabartay ( Caberda ), the Tatars captured 10,000 Circassians, and in that against yedukh (Bjaduk) and Aliyuk 40,000 or 50,000. Circassians then came to exchange their captured noblemen for 20 to 100 common slaves.

Let us conclude this rapid survey by asking the question whether the Ottoman society can really be defined as a slave society in the sense that its basic socio-economic structure was dependent for survival on an absolute control of labor or servile labor. We tried to show that the need for control of labor in Ottoman society varied in degree depending on the requirements of different segments of the society and economy. Slave in its classical sense, a person legally reduced to the nature of thing and subject to absolute possession and use by its owner, was something urgently needed in this traditional society in such enterprises as required large-scale, sustained and regular manpower -- not only for an imperial army and navy, colossal construction works, the elevation of large number of transport animals or large-scale agricultural production for the army or palace, but also in certain crafts, large estates and extended households in the society at large. However, even in these segments which, as a whole, constituted a limited area within the general socio-economic setup of the Empire, servile labor, with the exception of the konak (the extended household), and the large estates, disappeared over time especially from the end of the sixteenth century on when imperial laws were discarded and centralized power declined. We have already seen how in the second half of the fifteenth century the imperial government had great difficulty in maintaining its slave colony in the Khasskr. It became evident that servile labor was something extraneous within the Islamic agricultural system, and was exposed to constant and rapid erosion. The basic Ottoman organization of agricultural production rested upon the re'aya -- ciftlik system, that is to say, agricultural production was organized on the basis of small agricultural units (ciftliks) on the state owned lands placed in the possession of free peasant families (re'aya) -- Muslim or non-Muslimwho were under the sole obligation to cultivate it and pay taxes regulated by the state laws (kanun). They were free as defined by Islamic Law. No other person could force them to work or surrender the fruits of their labor without compensation. Even the restrictions such as the interdiction of abandoning the ciftlik land or the imposition of certain public services, such as mandatory work in mining or guarding the mountain passes, were not absolute, and were always mitigated by certain exemptions granted by the state. The imperial government was alert in preventing developments leading to the creation of personal ties of any sort over the reaya, and from the beginning made it a general policy to abolish serfdom or any other kind of personal tie over the rea-ya wherever it extended its rule. Of course, such a policy aimed first of all at weakening the power of the local lords in favor of the central authority. The Ottomans were aware of the fact that it served the expansion of their rule, and they utilized it as a propaganda device among the peasantry subject to serf- dom in neighboring countries.

The ciftlik-re'aya system was the dominant regime in the Ottoman Empire covering an overwhelming majority of the rural population along with most of the arable lands. It was not originally an Ottoman invention, but it was rather the re-establishment of an old system which had replaced servile labor as the major form of agricultural labor, economically the most advantageous and socially the safest under the demographic and economic conditions which prevailed during late antiquity. It can even be said that the Islamic Caliphate and the Eastern Roman Empire owed their enduring imperial existence to the fact that they embraced and became the champions of this regime which favored the continuance of independent peasant family units on the land.

At any rate what is known for sure is that this was the basic social and economic structure of the Ottoman Empire, carefully watched by the imperial government. Moreover, the Ottoman administration was ready to change the slave status of certain groups and eventually identified them with the re'aya when overall developments began to lead toward the disappearance of servile labor -- the difflculties in keeping a slave community within a majority of free peasants, the erosion of this community by frequent runaways and marriages with free men, and, perhaps the most important of all, the fact that slave labor became unprofitable under the new conditions of rising prices. But still, whenever private ownership under the Sharia stipulations was in question, especially on large estates and konaks any government disposition was excluded. It was only under the pressure from Western nations in the nineteenth century that the Ottoman government offlcially abolished slavery of any kind in its territories.