SLAVERY AND THE OTTOMAN-CRIMEAN

KHANATE CONNECTION

By

SHAWN CHRISTIAN BROYLES

Bachelor of Arts in History

Oklahoma State University

Stillwater, OK

2010

Submitted to the Faculty of the

Graduate College of the
Oklahoma State University
in partial fulfillment of
the requirements for
the Degree of
MASTER OF ARTS
May, 2014
 SLAVERY AND THE OTTOMAN-CRIMEAN

KHANATE CONNECTION

Thesis Approved:

Dr. Tonia Sharlach

Thesis Adviser

Dr. Lesley Rimmel

Dr. Yongtao Du

ii
 ACKNOWLEDGEMENTS

Thank you to Dr. Sharlach for guiding me through this difficult process when I had no
idea what was going on. To Dr. Rimmel who kept my work professional and of a high
standard. To my parents who have helped me get where I am.

iii
Acknowledgements reflect the views of the author and are not endorsed by committee
members or Oklahoma State University.
Name: SHAWN CHRISTIAN BROYLES

Date of Degree: MAY, 2014

Title of Study: SLAVERY AND THE OTTOMAN-CRIMEAN KHANATE

CONNECTION

Major Field: WORLD HISTORY

Abstract: This document discusses the economic connection between Ottoman Anatolia
and the Crimean Khanate within the context of the slave trade. It emphasizes the
consumption of slaves in Anatolia as a driving force behind the khanate’s consistent
raiding into Russian territory. The frequency of this raiding correlates with several
historical events and social practices which cause the demand for slaves to fluctuate.
Through the detailed analysis of this correlation, the Black Sea system is more effectively
described by a core-periphery relationship.

iv
 TABLE OF CONTENTS

Chapter Page

I. INTRODUCTION ......................................................................................................1

Thesis Statement ......................................................................................................1

Historiography and Sources .....................................................................................2

II. BEGINNINGS OF RELATIONS BETWEEN THE OTTOMANS AND THE

KHANATE ..............................................................................................................7

Historical Introduction of the Crimean Khanate ......................................................7

Ottoman Involvement in Crimean Khanate Politics ................................................9
Letters between the Khanate and the Porte ............................................................12
Slavery and Slave Raiding in the Crimean Khanate ..............................................16
Historical Data .......................................................................................................21
Discussion of the Data ...........................................................................................25
Conclusions ............................................................................................................26

III. SLAVE OWNERSHIP, BUREAUCRATIZATION OF THE TRADE, AND

REASONS FOR THE TRADE’S FLUCTUATION .............................................28

Slave Ownership in the Ottoman Empire ..............................................................28

Ottoman Taxation and Revenue from the Slave Trade ..........................................33
Ottoman Islamic Tradition, Slavery, and Manumission ........................................36
Changes in Agricultural and Price Realities ..........................................................38

v
Chapter Page

IV. CONCLUSIONS ...................................................................................................42

REFERENCES ............................................................................................................46

APPENDICES .............................................................................................................50

vi
 LIST OF FIGURES

Figure Page

1.1 23
1.2 24
1.3 51
1.4 52

vii
 CHAPTER I

INTRODUCTION

Thesis Statement

The Ottoman Empire and the Crimean Khanate had a very complex economic

relationship that lasted for over three centuries. This relationship revolved, in part,

around the capture, sale, and taxation of slaves who were taken on the Russian steppes

and beyond. This trade was centered between Istanbul and the major trade ports of the

North Black Sea system. From the trade’s official beginnings in the 1450s to its near-

death in the eighteenth century, it grew consistently. This growth is directly attributable

to social and economic realities in Anatolia. These include the empire’s

bureaucratization1 of the slave trade so as to increase tax revenue, the sultan’s subsequent

urging of the khan to increase slave raids, and Anatolia’s persistent need for slaves

caused by its people’s adherence to Islamic principles regarding slavery. These confluent

streams regarding the slave trade created a persistent and growing demand for slaves

from Anatolia from the trade’s bureaucratization in the 1450s to its end at the hands of

Russian Emperor Peter the Great. This pattern shows that the Ottoman Empire and the

1
By “bureaucratization” I am referring to the Ottoman practice of taxing new sources of revenue by
passing laws regulating the harvest and exchange of resources. Consequently, this creates a large paper
trail handled by an extensive system of Ottoman bureaucrats.

1
Crimean Khanate existed within a core-periphery relationship, with economic and social

realities in the Ottoman core influencing activities in the Crimean periphery.

Historiography and Sources

The Crimean Khanate was a breakaway state from the Golden Horde which, after

winning independence from the Horde, became an autonomous state in the north Black

Sea. As such, the Crimean Khanate adopted and continued the region’s well-established

economic system, most notably the slave trade. While there are several historians who

discuss the slave trade in Anatolia and the Crimea separately, few have discussed the

trade as a singular entity, and far fewer have focused on this exclusively through the lens

of the slave trade.

The historiography of the Ottoman Empire has a lengthy tradition, with a

veritable army of scholars dissecting one of the largest and longest-lived empires in

history. There are several notable scholars who have made grand contributions to the

field. Five of these authors are Halil Inalc ık, Donald Quataert, Saraiyah Faroqhi, Şevket

Pamuk, and Bruce McGowan.2

Inalcik and Pamuk stand out especially as both early modern and economic

historians, a group greatly aided by Ottoman bureaucratic practices that produced a

wealth of sources. Regardless of the criticisms placed on the empire by different

historians, the empire was very thorough in its bookkeeping practices. Such practices

2
Engin Akarli Deniz, “Ottoman Historiography,” Middle East Studies Association Bulletin 30, no. 1(1996):
33.

2
were necessary in running such a heavily centralized and patrimonial system. All assets

of the empire were ultimately owned by the sultan. Taxes on these assets, which were the

lifeblood of the empire and which provided a majority of the Porte’s budget, were

therefore thoroughly documented by trusted officials sent from the Porte. Inalcik has

made immense contributions to the field by capitalizing on these documents. His

translations of the Customs Tax registers at Caffa are a cornerstone of understanding

early Ottoman trade relations with the Black Sea economic system. He has translated

several maliye3 finance ministry documents held at the Topkapı Sarayi archives in

Istanbul. Inalcik’s summary of this wealth of economic data is in his two-volume edited

work, An Economic and Social History of the Ottoman Empire.4

Sevket Pamuk’s contributions are no less important. He has written several

articles and books which discuss the power and value of Ottoman currency. He has

compiled detailed and lengthy charts listing price changes of several commodities from

lumber to grain over several centuries.5 Both Pamuk’s and Inalcik’s works are the

primary starting points for any scholar wishing to explore early modern Ottoman

economics.

While all of these authors have made invaluable contributions to the field, they

are plagued by the problem of all Ottoman historians in that a majority of our available

3
The maliye was the Ottoman financial ministry, which was famous for its thorough bookkeeping practices
and extensive bureaucracy. All taxes collected by the empire were recorded in documents at the maliye.
For more information see Linda T. Darling, Revenue-Raising and Legitimacy: Tax Collection and Finance
Administration in the Ottoman Empire, 1560-1660: Ottoman Empire and Its Heritage (Boston: Brill
Academic Publishers, 1996).
4
Halil Inalcik and Donald Quataert, An Economic and Social History of the Ottoman Empire: vol. 1, 1300-
1600 (Cambridge: Cambridge University Press, 1997).
5
Sevket Pamuk, “Economic History of the Ottoman Empire.” University of Connecticut, 2000,
http://www.ottoman.uconn.edu/data.html, accessed 7 April 2014.

3
data is tied to the elite center of the empire. The further one moves away from the sultan,

the more sparse documentation on daily life in the early modern empire becomes.

This bias of information toward the Ottoman center of power slightly precludes

complete discussions of any topic that includes non-elites, especially slavery and the

Crimean Khanate, the foci of this paper.

While there are a great number of works on Ottoman slavery, they fall in line with

this data problem and focus almost exclusively on Janissaries and palace slaves, whose

lives are well documented; a quick online search will yield a wealth of popular history

books by authors such as Godfrey Goodwin or David Nicolle on the subject.6 Madeline

Zilfi discusses this elite slave narrative bias as a key component in the ignoring of the

story of female slaves in the empire, as women were not discussed in the context of

Ottoman military slavery.7

This has created a treatment of the early modern Ottoman slave trade that is less

negative than treatments of the later Atlantic slave trade and that tend to emphasize the

unique upward mobility of military slaves in the empire. Commentaries on non-palace

slavery-such as domestic slaves and agricultural slaves-are therefore largely from non-

Ottoman (usually European) sources. Mikhail Kizilov’s Slave Trade in the Early Modern

Crimea from the Perspective of Christian, Muslim, and Jewish Sources8 discusses the

problems of relying on these foreign sources, with their inherent confessional biases.

6
Godfrey Goodwin, The Janissaries (London: Al Saqi Books, 2006); David Nicolle, The Janissaries
(Oxford: Osprey Publishing, 1995).
7
Madeline Zilfi, Women and Slavery in the Late Ottoman Empire (Cambridge: University Press, 2010),
104.
8
Mikhail Kizilov, “Slave Trade in the Early Modern Crimea from the Perspective of Christian, Muslim,
and Jewish Sources,” Journal of Early Modern History, 11, no. 1/2, (2007): 1-31.

4
Major authors such as Omer Lütfi Barkan do not emphasize agricultural slavery in the

Ottoman Empire past the sixteenth and seventeenth centuries; in fact, Barkan contends

that agricultural slavery becomes the least common of slave activities.9 Determining the

percentage of slaves who engaged in agricultural activities is difficult because the entire

number of slaves in the empire at any given time is not known. Agricultural slavery,

however, was at least a major enterprise of borderland rejuvenation, as we shall see.

Again, the inherent problem of our historiography is the paucity of outside sources the

further one moves away from the center of power.

It is these outside sources, however, that become our primary vehicles for

discussing the periphery of the Ottoman Empire, especially the Crimean Khanate

borderland. Because the Crimean khans’ bureaucratic tradition was at best limited, we

cannot directly determine how many slaves they captured within a given year or see the

exact fate of these slaves. These numbers must instead be extrapolated from

administrative registers in the victims’ regions. Much of this computational work has

been done by Alan Fischer, who neatly used the work of eighteen Russian, Ukrainian,

and Polish historians to estimate the volume of khanate slave raids per year.10

These figures can be further complemented by the work of Inalcik, specifically

his translation of the records of customs tax revenues earned by the Ottoman Empire each

year on the slave trade. Combining estimates on the number of slave raids by the khanate

with estimates on the money made by the empire from taxing slaves entering through

9
Norman Itzkowitz, Ottoman Empire and Islamic Tradition (Chicago: University of Chicago Press, 1972),
88.
10
Alan Fisher, A Precarious Balance: Conflict, Trade, and Diplomacy on the Russian-Ottoman Frontier
(Istanbul: The Isis Press, 1999).

5
northern ports, and then comparing them with Ottoman policies and attitudes toward the

khanate, allows us to determine out the correlation of these factors.

The scholarship on the slave trade, thanks in great part to authors like Pamuk,

Inalcik, and Fisher, has created a great wealth of sources for discussing the economics of

slavery and the Crimean-Ottoman connection. I will show that the Ottoman Empire

became dependent on slavery, both on its actual practice as well as on the revenue made

from taxing it. Given that the primary source of slaves in the Ottoman economic sphere

was its northern borderland and the Crimean Khanate, the empire had to politically

position itself to the khanate in order to ensure that the khan remained both a semi-

autonomous subordinate and an ally.

Through the lens of the slave trade this relationship of power becomes readily

apparent. Fisher’s data on slave raiding shows a positive trend, which I discuss below.

As the Ottoman Empire became more dependent on slaves, demand for such slaves

increased even when there was economic instability in Anatolia. The factors behind this

dynamic are the main focus of this paper. From Islamic prescriptions ensuring that new

slaves were always in demand, to government projects requiring slave labor, the need for

slaves in Anatolia always trended upward. This dynamic is partially shown in both the

Ottoman budget and economic realities within Anatolia.

6
 CHAPTER II

BEGINNINGS OF RELATIONS BETWEEN THE OTTOMANS AND THE KHANATE

Historical Introduction of the Crimean Khanate

The cities on the Crimean Peninsula and the Sea of Azov served as the frontier

between the Ottoman Empire and Muscovy. This mutual borderland was heavily

populated and was a hub of commerce among Italy, Byzantium, Poland, Russia, and

Asia. Its chief city of commerce, Caffa11, was originally a Genoese trading colony

established in 1266 by an agreement with the Mongol khan. Within one century Caffa

became the center of an expansive and wealthy trade system. The Genoa State Archive’s

Diversorum Filze12, a collection of correspondence to the doge of Caffa dated from 1375

onward, mentions trading activities reaching from Trebizond to Lithuania.

The slave trade proved to be both a profitable enterprise for the Italians and a great

point of contention for the Mongol Horde. The Mongols took great offense at Italian

foreigners engaging in the kidnap and trade of Turkic peoples. This offense resulted in

11
The two most common names for this city are Caffa (Italian) and Kefe (Turkish). ‫ ;فك تلاا‬Eyālet-i
Kefê. Both names are used frequently in secondary literature on the subject. For the sake of continuity I
will use the Italian. The city is currently within Russian national borders and called Feodosiia, a slight
Slavicizing of the original Greek name.
12
S.P. Karpov, “New Documents on the Relations Between the Latins and the Local Populations in the
Black Sea Area, 1392-1462,” Dumbarton Oaks Papers 49 (1995): 33.

7
 the siege and destruction of Caffa in 1307, only for the Mongols to invite the Italians

back in 1312 under a new khan and allow the Italians to reestablish the slave trade.13

Despite these problems, the Genoese had already linked the Crimea and the slave trade to

a larger world economic system, setting a precedent for khanate exploitation of Italian

trade routes.

The beginnings of the Crimean Khanate were contemporaneous to troubles with the

Italian trading colonies. Once the Golden Horde had established an extensive Eurasian

land empire, many of its component parts began to integrate with local populations and

then eventually separate politically from the Golden Horde. The Crimean Khanate was

among these splinter groups that had established themselves in the northern Black Sea,

using the geographical isolation of the Crimea to vie for sovereignty from the Horde. As

Fisher has noted, the Crimea served as a refuge for “…unsuccessful aspirants to the

[Golden Horde] throne.”14 This seemed to be the perfect environment for launching a

rebellion against their former Mongol masters. After a lengthy war from 1420 to 1441,

the Crimean Khanate became an independent state with a complex and sovereign political

system. The Khanate’s political institutions were very much related to those of the

Golden Horde, while maintaining their own local practices.

While the khanate was politically separate from the Golden Horde, the former

maintained many core administrative and succession practices from the Horde. The

khanate was a loose confederation of nomadic tribes ostensibly ruled by the Giray clan

13
Caffa’s slave trade proved to be an ongoing flashpoint between the Italian trading colonies and the khans.
The city would be besieged again in 1348. According to an Italian witness to the siege, Gabrielle de
Mussis this was when the primary transmission of the Black Death to Europe began due to the Mongols
launching of infected corpses over the city walls.
14
Alan Fisher, The Crimean Tatars (Stanford: Hoover Institution Press, 1978), 3.

8
from the fifteenth century until its demise in the eighteenth century. Under the Giray clan,

the khanate respected the old political institutions of the Golden Horde despite the

former’s push for sovereignty. The Crimean khans were not absolute rulers and had to

respect the authority of the Karachi begs, the chieftains of the four most powerful clans, a

practice dating back to the Uzbek khans of the Horde.15 While it is difficult to ascertain

the extent to which the Karachi begs checked the khan’s authority, in the early years of

the khanate their signatures were required on any treaty signed by the khan in order to be

legitimate.16 The founder of the Giray dynasty, Haci Giray, was invited by other clan

leaders to become the Crimean khan and drew authority from his Ghengisid bloodline.

This gave him primacy among the clan begs.

This is the political context in which the khanate’s slave raiding and initial contact

with the Ottoman Empire occurred. From this point forward the khanate became a

semiautonomous vassal state providing Anatolia with the majority of its slaves. This

trade caused the two powers to develop a very complex but financially beneficial

relationship.

Ottoman Involvement in Crimean Khanate Politics

In 1475 Haci Giray’s son Mengli, who acceded to the throne in 1466, requested

Ottoman assistance in driving out the Italian presence on the peninsula. The Ottoman

15
Manz lists the four main clans as Shirin, Barin, Qipchak, and Arghin. She recognizes the ambiguity and
difficulty in identifying the exact clans as information on khanate political structure comes from a series of
correspondences with Muscovy from the years 1476 to 1515. These are, however, the four that appear
most frequently in the correspondence. For more information see Beatrice Manz, “The Clans of the
Crimean Khanate 1466-1532,” Harvard Ukrainian Studies 2 (1978): 282.
16
Manz, “The Clans of the Crimean Khanate 1466-1532,” 286.

9
Sultan Mehmed II used this opportunity to increase the Porte’s military presence in the

North Crimea and sent his admiral Gedik Ahmet Paşa.17 Subsequently the steppe

political system would become much more centralized with a definite Ottoman presence,

a process catalyzed and encouraged by the Ottoman Porte. This centralization benefitted

the Porte in that the empire could eventually control the khanate’s succession. The

empire’s ultimate interest in such control was predicated on Crimean stability, a steady

supply of slaves, and agricultural export.

As the khanate came under the influence of the Sublime Porte, the khan and the sultan

established a co-operative relationship. When the Ottoman Empire annexed a province, it

did so in stages, beginning with increasing tribute and concluding with the replacement of

local officials and administrative systems by the Ottoman Timar system.18 Mehmed II did

not attempt to impose this system on the khanate. Once Mehmed II assisted the khanate

in driving out the Genoese from the Crimean shore, he captured Mengli Giray, later

releasing him for accepting Ottoman suzerainty. Mengli Giray did not understand

himself as a vassal but as an Ottoman appointed official (tikme).19 Some scholars such as

Simon Sebag Montefiore suggest that the khanate’s bloodline came to be known as the

Ottoman’s “…third international tradition…”20 and that the khan himself was an Ottoman

official as Mengli Giray believed himself to be. Regardless of this relationship’s finer

17
Theoharis Stavrides, The Sultan of Viziers: The Life and Times of Ottoman Grand Vizier Mahmud Pasha
Angelović 1453-1474 (Leiden: Brill Publishing, 2001), 65.
18
Inalcik and Quataert, An Economic and Social History of the Ottoman Empire: vol. 1, 1300-1600), 13.
The Timar system was the Ottoman military-administrative system. It involved a large grant of land given
to an Ottoman official under the strict condition that regular revenues as recorded in the Tahrir surveys
were collected and given to the sultan. This system ensured the strength of the central government as
officials could be replaced or arrested for failing to meet the conditions of the original agreement. For
further discussion on the nature of this system and its departures from European feudalism, see Daniel
Goffman, The Ottoman Empire and Early Modern Europe (Cambridge: Cambridge University, 2007).
19
Halil Inalcik, “Struggle for East-European Empire 1400 – 1700: The Crimean Khanate, Ottomans and the
Rise of the Russian Empire,” Turkish Yearbook of International Relations, 21 (1995): 2.
20
Simon Sebag Montefiore, Prince of Princes: The Life of Potemkin (London, 2000), 244.

10
details, it is clear that the khan was held in higher regard than other foreign rulers. This

special relationship can be explained by the two states’ mutual sociolinguistic heritage.

The Crimean Khanate and Ottoman Empire were both Islamic states with Turkic

peoples, and they shared the common ideological heritage of the Islamic steppe warrior.21

In respect to their Islamic tradition, they both occupied similar goals in expanding the

dar-al-Islam22 through conquest. With these commonalities in mind, the sultans allowed

the khans to administer their own internal affairs and did not collect a regular tribute.

Instead, the sultan used them as a source of troops, but more importantly encouraged their

slave-hunting activities so as to have a constant source of labor for the Porte. This

relationship lasted for over two centuries.

The khans eased into their role as important members of a new economic order

centered on Istanbul. The warfare and trade in which they engaged played a crucial role

in Ottoman and Russian foreign policy for better or for worse. The khanate’s political,

economic, and subsistence structures evolved over time in order to facilitate Istanbul’s

need for slaves. Over time the khans became increasingly less autonomous, though they

were never fully vassals of the Ottomans.

In order to better understand this complex relationship, we can analyze letters

between the khan and the sultan from several points of time in their relationship. The

largest collection of these letters was translated into French by Mihnea Berindei and

21
The Ottoman Empire was certainly a multi-ethnic empire with subjects from several major language
groups including Indo-European, Semitic, and Turko-Altaic, but the language of government and elite
society was High Ottoman Turkish, a Turkic language with a wealth of Arabic and Persian borrow words.
The mutual cultural association of Islam and Turkic language certainly played into the special role
occupied by the khanate.
22
The “house” or “domain” of Islam. Frequently used in reference to all lands occupied by an Islamic
state, especially refers to lands contested with infidels.

11
Gilles Veinstein. The specific letters they translated were found in the Topkapı Sarayi

archives in Istanbul. These letters provide an insight into the power dynamic that existed

between the two states.

Letters between the Khanate and the Porte

The khan’s deference to the sultan is readily observed in both the introductions

and the content of their correspondence with each other. Several letters between the

Porte and the various khans were translated from the Topkapı archives. The letters date

from 1453 to the end of the eighteenth century. One of the earliest translated letters

between Khan Mengli Giray I and Mehmed II was written in 1479. Its tone of prostration

is seen in the opening lines:

Who had the splendor of the sun, the majesty of sovereignty, the freshness
of heaven and the glory of Adam who is raised as high as the stars, vast as the
sky, who possesses the help of Huma, the radiance of the sun, the justice of Kesra
[Chosroes], the generosity of hatem, the throne of Iskander, the good fortune of
Suleiman, who is the sovereign ruler of sovereigns of our time, which provides
safety and security, protects justice and the doers of good, in the shadow of God,
and who is the generous king, rich in generosity and justice, exterminator of
heretics and polytheists, the perfect vanquisher of the universe, rendering services
to religion, and assistant to Islam and Muslims, protector of the people of truth,
refuge [cave] of the people of Islam, defender of those who have found the right
path, slayer of transgressors, who propagates justice to the East and the West and
who manifests himself as the perfect man on both horizons.23

The pleading nature of this letter’s introduction is obvious, invoking classical

images of power relevant to both Islamic and Ottoman Turco-Persian heritage. What is

even more curious is the khan’s reference to the sultan as his “frère” (brother). This

23
Bennigsen et al., Le Khanat de Crimée dans les archives du Musée du Palais de Topkapi (Paris: The
Hague, 1978), 41, trans. Shawn Broyles; a letter from the Khan Mengli Giray to Mehmed II Fatih.

12
suggests a closeness with co-religionists not shared in the correspondences with the rulers

of other nations. This could be further debated as there is a tradition of Near Eastern

kings using familial identifiers for each other, but certainly not for European rulers.

While this letter is certainly flattering and aggrandizing to the sultan, it still does

not denote Ottoman influence in Crimean affairs. Most letters from this era begin with

similar addresses to the sultan. In the same letter Mengli Giray then attempts to resolve a

matter with the sultan wherein one of the sultan’s naval commanders has taken military

action against his men around Caffa. The khan pleas with the sultan to act as a brother

and reminds him that “Kefe est votre ville tributaire, mais elle est aussi notre résidence”;

Caffa is your tributary/dependent, but it is also our residence.24

Ottoman authority over the khanate increased alongside income from major slave

trading cities such as Caffa. By 1547 the Ottoman central government was regularly

sending military orders to the khan. One such order commanded the Khan Sahib Giray to

immediately strike back against Russian invaders at Uzak. The letter ordering the strike

employs authoritative language:

No delay, no delay can be tolerated. Act with diligence and depart on a

relentless campaign. In addition, we have ordered the bey of Caffa to chastise the
heretic troublemakers.25

The Crimean Khanate played an integral role in Ottoman military policy toward

Russia, but posturing to the Porte’s northern neighbor was only one of many Ottoman

24
Ibid.
25
Ibid., 131.

13
concerns that the khanate handled. The Crimean khans were also expected to honor and

protect Ottoman trade agreements. A 1574 letter from Selim II to Khan Devlet Giray

demands that the khanate provide escort and protection for a functionary merchant of the

Porte named Mikhail.

Mikhail, one of the merchants of our imperial court, bearer of a sovereign

and majestic order, model of the notable men of the Christian nation, charged
with the purchase of sable and other merchandise destined for our imperial court,
is sent to the land of Muscovy. This mission is very important and necessary. He
must be able to arrive in complete security to the said land. When the said
merchant arrives bearing our joy-dispensing letter, you—observing the customs of
former times and in conformity with the loyalty and fidelity that you have always
shown to the sublime Porte—must do everything necessary to accompany the said
merchant with an appropriate escort, so that he might, safe and sound, arrive there
and return in stages to our blessed threshold.26
By the middle of the sixteenth century, the Ottoman Porte’s interactions with the

khanate resembled a relationship with a subordinate client state rather than with a

mutually respected co-religionist state. While there are many factors contributing to the

increasing Ottoman influence, such as a family and tribal disunity weakening the position

of the khan, Ottoman need to increase revenue and take advantage of the weakness of the

khan was an integral part of Ottoman policy and influence in the Crimea.

Despite the special relationship that existed between the two powers on an

ethnolinguistic and confessional basis, the sultan and the khan were occasionally at odds,

with the Porte constantly trying to bring the khanate within its bureaucratic structure.

There are several instances in which the khan outright refused aid to the sultan, although

these refusals were mostly isolated to the earliest periods of Ottoman-Khanate

26
Ibid., 138.

14
interactions.27 As their relationship continued, the Khan often would not refuse but

would demand some sort of financial incentive to assist. By the end of the sixteenth

century, most correspondence, such as the imperial order to Sahib Giray was strictly

followed. This was likely due less to a fear of Ottoman reprisal than to economic

integration of the slave trade. It was mutually beneficial, especially in the late stages of

the Ottoman-Crimean relationship, to maintain the revenue generated by slave raids. The

bureaucratization of the slave trade and the volume of its subsequent revenues caused not

only more demand for slaving activity to supply Anatolia, but created the impetus for

Istanbul to exercise increasing amounts of authority over the semi-independent Crimean

Khanate.

These letters demonstrate that the power dynamic between the two, while

complex, favored the sultan. However, the letters show only deference and are not

damning evidence of Ottoman involvement in khanate slave trade policy. There are three

letters from the late sixteenth century from the sultan, who explicitly encouraged the khan

to increase the frequency and intensity of Tatar raids into Russia.28 Among these is a

specific decree issued in 1586 under Murad III to the governor of Caffa ordering the

latter to relax certain trade formalities in order to increase slave supply.

Perhaps the most convincing example of increasing Ottoman control in the

khanate came at the time of Mohammed Giray and his revolt against Ottoman authority
27
During the first Ottoman-Safavid War (1532-1555), Suleiman the Magnificent requested that the
Crimean khan provide assistance in the form of 50,000 horsemen. This request was declined. On some
occasions the khans asked for exorbitant fees to participate, Alan Fisher, A Precarious Balance, 31.
28
Alan Fisher, “The Sale of Slaves in the Ottoman Empire: Markets and Taxes on Slave Sales, Some
Preliminary Considerations,” Beseri Bilimler-Humanities, Vol. 6 (1978); Miihimme Defter LXII, 119
(1586), 52; also Miihimme Defter VI, 624 (1564), pp. 29-30; XXIII, 248 (1573), 122; XXIII, 295 (1573),
145; and LXVIII, =^116 (1590 - 1), 60, all to the governor of Kefe to encourage the Çerkes to supply more
slaves; LXXII, 193 (1593-4), 101, to the governor of Kefe to encourage the Tatars to do the same. As
quoted on p. 171. I have tried to obtain these documents but it has proven to be very difficult.

15
in 1583. His lack of willingness to fulfill his predecessors’ role of military support in the

Persian campaign, along with his desire to keep the throne in his family, ran counter to

the sultan’s view of this role. This led directly to Mengli Giray’s overthrow and

deposition in 1584 by the Ottoman sultan, a clear statement of power.29 Anti-Ottoman

factions continued to exist within the khanate but the sultan always kept a firm hand on

politics in the Crimea.

These letters in conjunction with hard economic evidence suggest deliberation

on behalf of the Porte in influencing Crimean affairs. This represents a “core” state

exploiting and eroding sovereignty in a “client” state for resources.

The following section deals with the economic evidence of this exploitative

dynamic. It discusses the prevalence of slave raiding in the Crimea. It also couples a

graph showing an increase in slave raids in the Crimea with data discussing this increase

as existing parallel to Ottoman economic involvement in the slave trade.

Slavery and Slave Raiding in the Crimean Khanate

While the finer details of The Crimean Khanate political system are difficult to

pinpoint, the khanate’s main source of income is not. The Crimean Tatars were infamous

throughout Kazan’, Astrakhan’, Poland, and Russia as the chief agents in the region’s

slave trade. This trade was a key socioeconomic institution within the Crimean Khanate.

This study will focus on the slave trade between the khanate and the Ottoman Porte and

29
C.M. Kortepeter, “Āāzī Girāy II, Khan of the Crimea, and Ottoman Policy in Eastern Europe and the
Caucasus, 1588-94,” The Slavonic and East European Review 44, No. 102 (1966), 143.

16
how changes in this institution spurred greater socioeconomic changes within the Black

Sea system. Specifically, it will describe the demand for slaves created in Anatolia and

how that demand affected Crimean Khanate policy and sovereignty.

The slave trade in the Black Sea region existed long before the predations of the

khanate, all the way back to Byzantium.30 The fading influence of the Byzantines,

especially after the Battle of Manzikert (1071), allowed groups like the Seljuk Sultanate

of Rum to build their wealth on the region’s well-established slave trade.31 As mentioned

previously, this trade was expanded from Anatolia to the Baltic by the Genoese trading

colonies of the Crimea.

Under the leadership of the Giray dynasty, however, the khanate continued this

tradition with a new fervor that would leave a lasting social and economic trauma on the

region. The devastation their raids wrought on Poland and Ruthenia emptied the

countryside and created a great amount of unrest among local Christian populations,

resulting in a series of Cossack uprisings in the mid-seventeenth century.32

Several Russian, Ukrainian, and Polish folk songs from this period illustrate the

impact of the raids on common people in these regions.

The fires are burning behind the river—

The Tatars are dividing their captives.

30
Fisher, A Precarious Balance, 28.
31
C.M. Kortepeter, “Ottoman Imperial Policy and the Economy of the Black Sea Region in the Sixteenth
Century,” Journal of the American Oriental Society 86 (1966): 87.
32
Dariusz Kołodziejczyk, The Ottoman Survey Register of Podolia ca. 1681 (Boston: Harvard University
Press, 2004), 4.

17
 Our village is burnt and our property plundered.
Old mother is sabred
And my dear is taken into captivity.
(a Ukrainian folk-song)33

Folk songs and tales such as these were common. Dariusz Kołodziejczyk points out

that as late as 1948, a series of local village history projects produced an overwhelming

number of tales about the Tatar horror.34

Once the new captives were taken from their homeland, differing fates awaited them.

In general slaves were destined to become the “plough and scythe”35 of their masters in

agricultural endeavors. Many became galley slaves who were purchased in great

numbers by the Porte.36 Sometimes slaves became something more devious. Evliya

Celebi’s Sayahatname (record of travels during the mid-seventeenth century) discusses

the practice of both male and female sex slavery in Anatolia and Tatary.37 It is important

33
Adrian Kashchenko, Opovidannia pro slavne Viis’ko Zaporoz’ke nizove, (Kiev, 1992), 29. As quoted in
Mikhail Kizilov, “The Slave Trade in the Early Modern Crimea From the Perspective of Muslim, Jewish,
and Christian Sources,” Journal of Early Modern History 11 (2007): 1.
34
Franciszek Kotula, “Warownie chlopskie XVII w. w ziemi przemyskiej i sanockiej” [Peasant strongholds
in the districts of Przemysl and Sanok in the seventeenth century], Studia i materialy do historii
wojskowosci VIII, pt. 1 (1962), 73-149. as quoted in Dariusz Kołodziejczyk, “Slave Hunting and Slave
Redemption as a Business Enterprise: The Northern Black Sea Region in the Sixteenth to Seventeenth
Centuries,” Oriente Moderno 25 (2006): 149. Mikhail Kizilov’s discussion on “The Moral Side of the
Crimean Slave Trade” points out the difficulty of determining the well-being of slaves as many accounts,
mostly Christian, have an obvious bias against painting the Tatars in a humane light.
35
Archiwum Glowne Akt Dawnych: Archiwum Koronne Warszawskie, Dzial Tatarskie, k. 61, t. 135, no
277 (1661, Polish) as quoted in Mikhail Kizilov, “The Slave Trade in the Early Modern Crimea,” 14.
36
Eizo Matsuki, “The Crimean Tatars and their Russian Slave Captives: An Aspect of Muscovite-Crimean
Relations in the 16th and 17th Centuries,” The Mediterranean Studies Group Hitotsubashi University 18
(March 2006): 177. Matsuki writes, “In 1576 the Ottomans navy staffed twenty galleys with Slavic slaves,
purchased 6000 slaves in 1579, and 4000 in 1590 as well.”
37
R. Dankoff & K. Sooyong, An Ottoman Traveler: Selections from the Book of the Travels of Evliya
Celebi (London: Eland Publishing, 2010), 249. There were many different forms of slavery in the Ottoman
Empire and as such the subject cannot be addressed in a fashion similar to that of western slavery. For
more complete discussions on the status of various slaves in Ottoman society during this period, see:
Yvonne Seng, “Fugitives and Factotums: Slaves in Early Sixteenth Century Istanbul,” Journal of the
Economic and Social History of the Orient 39 (1996), 140; Alan Fisher, “The Sale of Slaves in the Ottoman

18
to note that Islam and the Islamic state were not always complicit in these practices. The

Qur’an is very clear about Islam’s position on slavery and placed restrictions on who

could be made a slave, specifically non-Muslims captured in war.38

The Qur’an simultaneously emphasized emancipation as a charitable act, a practice

which was common and documented in countless Ottoman sources. It is in fact the

Qur’an’s encouragement of the emancipation of slaves that caused a consistent demand

for non-Muslim slaves from abroad and fed the need for increased slave raids into

Christian lands throughout the sixteenth and seventeenth centuries.

The economic impacts of Tatar slave raiding were just as scarring as its social effects

but were equally as profitable for the khanate. The raids were so severe in Muscovy by

the mid-seventeenth century that a separate fund39 was created by the Ulozhenie (law

code) of 1649 to pay the ransoms of slaves captured by the Tatars.40 The amount of

money exacted from those paying ransom can only be estimated but it was no doubt

enough to be considered an integral part of the khanate’s economy. A letter written in

1533 from a khanate official to the sultan calls slave raids into Christian lands the source

of the Crimean people’s prosperity.41

Empire: Markets and Taxes on Slave Sales, Some Preliminary Considerations,” Beseri Bilimler-Humanities
6 (1978): 151.
38
The Qur’an has several verses on the practice of slavery and on the divine benefits of emancipating one’s
own slaves. For more information reference these suras: Al-Baqarah 2:177; Al-Balad 90:1-15; An-Nur
4:32-33; An-Nur 2:221.
39
Fisher, A Precarious Balance, 28.
40
A.G. Man'kov et al., ed., Sobornoe Ulozhenie 1649 goda: tekst, kommentarii (Leningrad: Izdatel'stvo
"Nauka," 1987), 27-28, 172-73; Russia; and Richard Hellie, The Muscovite Law Code (Ulozhenie) of 1649
(Irvine, CA: Charles Schlacks, Jr., Publisher, 1988)
41
Bennigsen et al., Le Khanat de Crimée , 121-25 as quoted in Halil Inalcik, “The Khan and the Tribal
Aristocracy: The Crimean Khanate under Sahib Giray I,” Harvard Ukrainian Studies 3-4 (1979-1980):
452.

19
 Khanate slave raids into Christian lands occurred annually from the early fifteenth

century until the end of the seventeenth century, ceasing almost entirely due to the efforts

of Tsar Peter I and his conquest of Azov.42 Once the khanate’s political structures were

taken under suzerainty by the Ottoman Porte in 1475, Crimean slave raiding and trading

would be permanently defined by Ottoman demand.

The extent of this demand can be partially measured by the averages of populations

taken by slave raids over the time period of 1468 to 1688. These figures show the

enthusiasm of the Crimean khanate slaving machine over time, an enthusiasm directly

affected by Ottoman need for slaves. If we compare the flux in the number of slaves

captured per year to contemporary economic realities in Anatolia, we can discern more

complex themes concerning the relationship of the khanate to the Ottoman Porte and

place the two powers appropriately within a core-periphery model. Before we look at

slave raiding data, it is necessary to emphasize the extent to which slavery permeated

Ottoman and Crimean society.

Khanate participation in the slave trade relied largely upon and was perpetuated by

the demand for slaves created by the Ottoman central government. What is interesting

about the khanate slave trade is its nature as a purely “economic”43 activity that did not

seem to have any purpose other than to perpetuate itself and profit those involved. The

lucrative profit of this industry was generated by its primary customer, the Ottoman

central government, which purchased roughly seventy percent of Crimean imported

slaves from the largest Black Sea slave market in Caffa.44 In turn, this industry’s primary

42
Fisher, A Precarious Balance, 45.
43
Matsuki, “The Crimean Tatars and Their Russian Slave Captives,” 175.
44
Fisher, Alan. A Precarious Balance, 36.

20
customer would have quite an effect on the ventures of Crimean Khanate elites. Cross

referencing data on slave raids and primary source documents on Ottoman policy allows

us to define this effect.

Historical Data

Domination of the Black Sea slave trade had many benefits for those powerful

enough to do so. From the mid-fifteenth century until the nineteenth century, the

Ottoman Empire reaped the extensive financial benefits of the slave trade, albeit at the

cost of fomenting war with its neighbors. This trade expanded greatly within the

centuries during which it was under Ottoman “management,” a trend which is evidenced

by historical data. The Ottoman slave trade existed to fill a necessity within the empire

but also existed as an enterprise growing for its own sake. The more slaves that passed

through the empire’s port cities, the more money was made off of their trade, whether the

slaves’ final destinations were in the empire or not. This need to increase both actual

slave populations in Anatolia (initially), combined with the need to collect increasing

revenue from their trafficking, explains the immense demand created, a demand which

could only be readily supplied en masse by the Crimean Khanate. The following data is a

representation of the average number of slaves taken per annum by the khanate during

this period. Its relevance operates under the assumption that Ottoman demand for slaves

was a primary driving force behind khanate slave raiding.

While the use of exact figures in historical research is problematic, it is important to

analyze quantitative data to observe general trends in history. Figure 1.1 is a graph of the

21
average recorded number of slaves taken by Tatars from Poland, Kiev, Muscovy, L’vov,

South Russia, Galicia, and Ukraine. The “spikes” in the graph all correlate to events of

major expansion and/or transformation of the Ottoman system in its center.

22
Figure 1.1

23
 45
Figure 1.2

45
I compiled this data from figures in A. Fisher’s A Precarious Balance. I omitted yearsrs that had an
ambiguous figure such as “thousands” or no number at all. Fisher compiled this data from several Russian
and Polish historians: Neil Kressel; Khrisanf Petrovich Iashchurzhinskii; Mykhailo Hrushevskii; M. A.
Alekberli; M. N. Berezhkov; Eudox
Eudoxiu
iu de Hurmuzaki; Isaak Massa; V. A. Golobutskii; Georgii Koniskii; A.
A. Novosel’skii; Maurycy Horn; D. I. Bagalei; Wiesław Majewski; Sieur de Beauplan; V. D. Smirnov;
George Vernadsky; Bohdan Baranowski; Tiapkin; and the famous Ottoman traveler Evliya Celebi. Cel Fisher
notes that many of the numbers are unreliable. Many parties, especially Polish and Russian historians, had
an interest in inflating the numbers to emphasize the impact of Tatar raids. This makes obtaining exact
figures a pointless task. However
ever my argument requires the observation of macroeconomic trends rather
than microeconomic events. This graph likely provides an accurate picture of general size trends in slave
raids over its time frame and is thus very helpful in discussing the largest and most sweeping changes
within the empire as they relate to creating commodity demand in the periphery. Thanks to David Phillip
Arthur Craig, a graduate statistician at Oklahoma State University, who provided four different graphs with
a full write up on the statistical methods used. He utilized a method known as Observation Oriented
Modelling. In summary, without the 400,000 slaves outlier, the graph trends slightly negative yet 65.05
percent of the data suggests a positive trend. The graphs provide
providedd are the most relevant given the focus of
this paper.

24
 Discussion of the Data

The two graphs provided are separate correlations of the same data with positive

trend lines. The reason for presenting the data in this way is to emphasize the tendency

of slave raiding to increase over time regardless of mathematical outliers. The large

400,000 outlier is solely responsible for the positive trend in the first graph; without it the

overall line is slightly negative. This is not to say that slave raiding as an average

decreased over time, which is not true, but to say slave consumption and therefore raiding

had an upward tendency which was subject to social and economic realities. The

statistical write-up at the end of the document confirms that 65 percent of the data points

to a general increase in slave raiding. This general increase is shown in the second graph,

which splits episodes of slave raiding before and after a time of great economic distress

in Anatolia which caused overall consumption to decrease. However, after this period

consumption continues upward, as is shown by the second trend line in the red dots.

There is a clear positive trend that reflects the reality of consistent increase in both

slave raids in the borderlands and slave consumption in Anatolia. One would expect the

number of slaves required to meet Ottoman needs to eventually level out once those

needs were met. This is, however, not the case. Due to several factors within Ottoman

society, the need for slaves always fluctuated and almost always had a tendency to

increase. This trend is directly attributable to three factors. The “dip” in this graph

around 1600 likewise has an explanation which reflects economic realities in Anatolia.

First, the Ottoman bureaucratization of the slave trade allowed the Porte to increase

its tax revenue from the trade. By looking at the immense amounts of wealth and the

25
Porte budget’s subsequent dependency on the trade, we can see why the empire went to

such lengths to have a beneficial and later controlling relationship with the khanate.

Next, the prescriptions of Islam on slavery created an atmosphere in Anatolia wherein

the slave population constantly needed to be replenished. Manumission was both a pious

and common act. Despite this, the need for slaves remained high and the main source of

those slaves in both government and private enterprises was the khanate trade.

Finally, the low point in the graph shows that these three factors contributed to the

upward trend in figure 1.1 and were the primary reasons for the intensification of slavery

in Anatolia and Ottoman involvement in Crimean politics. This low point coincided with

the height of the Celali revolts, a series of internal conflicts in Anatolia which brought

government activity to a screeching halt, and subsequently the empire’s purchase of

slaves from the Crimea. More importantly, this period seems to have been initiated by

the events of 1571 wherein the Crimean Khanate sacked Moscow. The number of

captives taken listed in figure 1.1 and 1.2 are 100,000, which marks it as a significant

event. The Russian reaction in the following year resulted in the khanate’s defeat at the

Battle of Molodi. This study does not focus on the events at Moscow or Molodi due to

the thesis’s emphasis on Crimean-Ottoman relations. Further study into these events and

using with the current data have the potential to provide new and important insight to the

mechanics of the Black Sea system.

Conclusions

The Crimean Khanate utilized a well-established trade system to prosper once it

separated from the Golden Horde. This trade grew to become the khanate’s most

26
profitable industry beginning in the mid-fifteenth century, so much to the point that the

khanate was considered a “slave society.”46 Once the khans separated from the Golden

Horde and created their own identity, they resumed and grew the region’s slave trade but

only to the extent of fuelling their own need for slaves and supplying European traders

with slaves. Once the khanate became a client state of the Ottoman Empire, the slave

trade increased dramatically. The position of the khan quickly became subject to the

sultan’s political needs. As time wore on and slave raids increased, so did Ottoman

income. This situation, which mostly benefited the sultan, was deliberately created by

Mehmed II and encouraged by later sultans. An investigation into the revenue gained,

based on the growing slave trade, will further illuminate the idea of the Porte creating a

more profitable economic situation in the Crimea.

46
James Walvin, “Slavery,” Oxford Companion to Black British History, May 2009,
http://www.oxfordaasc.com.gate.lib.buffalo.edu/article/opr/t240/e380, accessed 5 June 2014.

27
 CHAPTER III

SLAVE OWNERSHIP, BUREAUCRATIZATION OF THE TRADE, AND REASONS FOR

THE TRADE’S FLUCTUATION

As an important foreword to this chapter it is necessary to outline the meaning of

demand, which is an integral part of this thesis’s argument. Demand, for slaves in this

instance, is the economic term which references the extent to which a good fulfills the

desire of the consumer. In this case the demand of slaves is determined by how useful

they were in agricultural and domestic enterprises owned by the Ottoman consumer of

slaves. As slave populations decreased in Anatolia, demand for them became greater

because of the valuable purpose slaves served. Demand in this context fulfilled the need

to replenish slave populations. The prevalence of slave ownership in the Ottoman

Empire creates this demand, which tended to increase.

Slave Ownership in the Ottoman Empire

In order to emphasize the empire’s deep-seated interest in slaves from the Crimea,

it is necessary to emphasize the extent to which slavery permeated Ottoman society.

Slave ownership was very prevalent in Ottoman society. While slaves could be found

performing just about any duty that a free Ottoman subject could do, they were most

often near the pinnacle of power in the Porte and consumed in large numbers by the

political elite. The military (askeri) and ruling class more specifically were the most
28
 prominent consumers of slaves. Almost all palace servitors and military

personnel closest to the sultan were or continued to serve as slaves. It was not

uncommon for the court’s highest functionaries such as the grand vizier to have hundreds

of slaves. Rüstem Paşa, the grand vizier under Suleiman the Magnificent, is recorded to

have owned seventeen hundred slaves upon his death.47 The sultan had thousands of

Janissaries in his service at any one time. However these slaves came from the peculiar

Ottoman institution of Devşirme48 rather than a harvesting of the steppe. Slavery

becomes less prevalent in Ottoman society the further away one moves from the court.

Records left behind by the wealthy elite in Edirne (modern Adrianople) show that 41 out

of 93 recorded estates owned slaves, or 140 individual slaves total.49 By the mid-

seventeenth century almost all palace functionaries were slaves of one type or another.50

The lower strata of society employed slaves but certainly not in the volume in

which their social superiors did. Slaves occupied all areas of peasant society from rural

agriculture to urban crafts and trades. Legal estate records from Uskudar called sicilli

defterleri list the slaves owned by the deceased and their value.51 Many of these estate

records were kept by farmers and artisans who often only needed around two slaves for

the daily running of their estate. The defterleri discussed by Yvonne Seng lists 153

47
Steven Murray, Islamic Homosexualities: Culture, History, and Literature, (New York: New York
University Press, 1997), 178.
48
The Devşirme was an Islamic tax of the Ottoman state aimed specifically at non-Muslim populations. It
took one out of every forty boys and raised them near the Porte to be the sultan’s own guard and check
against the Turkic nobility. The practice began under Murad I but tapered off significantly in the sixteenth
century. For more information on the Devşirme, see Halil Inalcik and Donald Quataert, Social and
Economic History of the Ottoman Empire vol 1, 1300-1600, 283.
49
Halil Inalc ık, “Servile Labor in the Ottoman Empire,” 24-53 in Abraham Ascher, Tibor Kiraly, and T.
Halasi-Kun, (eds.), The Mutual Effects of the Islamic and Judeo-Christian Worlds: The East European
Pattern (New York: Brooklyn College, 1979) 25-43.
50
Rifaat Ali Abou-El-Haj, “The Ottoman Vezir and Paşa Households 1683-1703: A Preliminary Report,”
Journal of the American Oriental Society 94, No. 4 (Oct. - Dec., 1974), 440.
51
Istanbul Seri'ye Sicilleri: Uskudar Series 6, volumes 3 and 6.

29
estates, only 11 of which employed slaves. In the case of most businesses or farms the

number of slaves per estate was equally small. The number of slaves for one Musa’s

estate of 25,550 akçe only employed two slaves.52 Madeline Zilfi states that at the end of

the sixteenth century, roughly 20 percent of Istanbul’s population consisted of slaves or

freedmen with half of those employed in Bursa or Istanbul’s artisan and silk industries.53

Slaves came from many different places on the borderlands of the empire. The

main support we have for the khanate as the primary supplier of slaves to the empire is

the overwhelming presence of slaves of Slavic and Circassian origin within local court

records.54 A slave’s origin could possibly influence their fate. Slave holders

overwhelmingly demanded a flow of white Slavic slaves, while the “black” slave trade

primarily fed the needs of Egypt and the Mediterranean.

The status and quality of life for the slaves themselves varied, depending on their

function and masters. The most common legal mechanism applied to slaves was called

the mukatebe contract.55 During the slave’s term of service, he or she would negotiate on

a price to buy their freedom. This was encouraged by the Islamic religious community,

as manumission was considered a pious act. Nevertheless, slaves were not legally in

possession of their labor or bodies and were treated as property. Their appearance in

estate registers right next to common household items attests to this widely-held

52
Seng, “Fugitives and Factotums,” 142.
53
Zilfi, “Women and Slavery in the late Ottoman Empire,” 130.
54
Yaron Ben Naheh, “Blond, tall, with honey-colored eyes: Jewish ownership
of slaves in the Ottoman Empire,” Jewish History 20 (2006): 317.
55
Inalc ık, “Servile Labor in the Ottoman Empire,” 24-53.

30
treatment. The only exception made between regular property and slaves was the legal

Islamic distinction between property “without a voice” and property “with a voice.”56

This is a significantly darker picture of Ottoman slavery than has been depicted in

previous approaches, which have emphasized the immense opportunities open to slaves.

These positive arguments often cite the many viziers and military commanders of low

birth who rose to prominence in the Ottoman political structure. These experiences are

unfortunately in the minority. These experiences also usually occurred within the context

of the sultan’s Devşirme tax, which could often be beneficial to slaves. A majority of

slaves, however, were bought in large city center markets and therefore did not have

access to the much-celebrated upward mobility of the Ottoman kul.

While largely confined to the elite, slave ownership was practiced in all layers of

Ottoman society. Farmers and artisans alike employed small numbers of slaves bought

from their local markets. The upper-class and military employed thousands of slaves as

palace functionaries, workers on their estates, and soldiers. These slaves were acquired

through a combination of market purchases and the Devşirme tax.

While there are no reliable estimates on the exact number of slaves within the

entire empire at any one time, there are palace and court records, taxation registers, and

personal accounts that support this thesis’s contention of an expanding slave trade over

time. According to the data, this demand increased exponentially over just a century.

56
Ibid., 139. Sahillioglu, Halil 1985; mali samit and mali natik, respectively. Property with a voice
included cattle alongside humans. For more on the Islamic legal concepts of property, see Itzkowitz,
Ottoman Empire and Islamic Tradition, 88.

31
 Mehmed II’s ascension to the throne began this trend by expanding the sultan’s

slave-military compliment to 15,000 soldiers. This number would reach 100,000 by

1609.57 This trend unknowingly created an atmosphere of dependency on slave labor that

would redefine the Ottoman Empire’s relationship with coreligionists across the Black

Sea.

Ottoman law regarding the reclamation of land also helped to encourage a

demand for slaves. Mehmed II’s plans for expanding and revitalizing Istanbul from 1453

to 1480 involved the resettlement of 163 villages to resume interrupted agricultural

activity. These settlements, known as Khasslar58, were populated by slaves bought from

Mehmed II’s re-established slave trade in the north.59

General consumption on all levels of society, government projects that utilized

slave labor, the massive Ottoman military machine, and agricultural works in Anatolia all

point to an extremely large demand for slaves in Anatolia and provide reasonable

grounds to point to Anatolia as a driving factor in demand for the Crimean Khanate’s

slaving activities. By looking at Ottoman finances regarding the trade, this becomes

more apparent.

When we look at the vast amounts of money made by the Porte from this trade, it

is not difficult to see why the sultans so heavily encouraged the practice and thus created

demand and increased slave raiding.

57
Inalcik, “Servile Labor in the Ottoman Empire,” 24-53.
58
Khasslar lands were lands owned by the state and settled with slave labor as an attempt to create instant
revenue for the Porte. Mehmed II’s revitalization of Istanbul is the greatest example of this phenomenon.
For more information regarding Khasslar see Halil Inalcik, “Servile Labor in the Ottoman Empire,” 40.
59
Ibid, 31.

32
 Ottoman Taxation and Revenue from the Slave Trade

The Ottoman state had been collecting taxes on slaves since the fourteenth

century. Murad I established the law of pencik, which provided the sultan one fifth of all

slaves taken from the enemy in a time of war, with a tax of 25 akce per slave.60 Even

before the official bureaucratization of the slave trade, that sultan had realized the

importance of taxing the slave trade. Murad I’s vizier noted specifically how detrimental

it would have to have been to have foregone this tax. The tax on slaves vastly increased

by the mid-fifteenth century with the creation of the first state-administered slave market.

The first four points in figure 1.1 represent a steady increase in slaving activity from

the late 1460s onward. The first official slave market61 (esir pazari) was established by

Mehmed II in Istanbul in the 1460s.62 The first “major” Tatar slave raids occurred in

1468.63 While the Ottomans engaged in this trade from 1399 onward, this slave market

was the first example of the Ottoman government bringing the practice under the Porte’s

regulation as a source of state revenue. From the trade port of Caffa alone the Ottoman

Empire made over one million akçe in customs tax64 from the slave trade during the years

1487 to 1490. By 1526 to 1527, that revenue had increased to 2.5 million akçe.65 This

60
Kate Fleet, “Tax Farming in the Ottoman Empire,” The Medieval History Journal 6 (2006): 253.
61
Fisher, “The Sale of Slaves in the Ottoman Empire: Markets and Taxes on Slave Sales, 151.
62
Ibid, 171.
63
Idem, “Muscovy and the Black Sea Slave Trade,” Canadian-American Slavic Studies 6:4 (1972): 579.
64
Inalcik, Economic and Social History of the Ottoman Empire, 280.
65
Két tárgyalás sztambulban. Hyeronimus Laski tárgyalása a tӧrӧknel János király neveben.
Habardanecz János jelentese 1528 nyari sztambuli tárgyalásairol [Two Sets of Negotiations in Istanbul.
The Talks of Hyeronimus Laaski with the Turks in the name of King John. The Report of Johannes
Habardenecz on the Negotiations in Istanbul in the Summer of 1528]. Published by Gábor Barta. Trans.
by Gábor Barta and Jozsef Kun. Budapest, 1996, 138. As quoted in Pál Fodor and Géza David, Ransom
Slavery along the Ottoman Borders (Leiden: Brill Publishing, 2007), XIII. The figure listed in this source
is 30,000 gold ducats which I calculated and converted into akçe for the relevant years. These calculations

33
steady increase, which began in 1468, correlates with the establishment of Mehmed II’s

Istanbul market and the entrance of the Ottomans as customers and tax collectors of the

Crimean slave trade.

This increase is no coincidence. As far as our primary sources tell us, slave raids by

the khanate before 1475 were few in number when compared to those of the early

sixteenth century. Sources from the Latin and Greek merchant communities in the

Crimea indicate that while still profitable, the slave trade declined significantly after the

conquest of Constantinople in 1453 and the following regional turmoil, demonstrating

how Constantinople/Istanbul had been a major node in a well-established slave trading

system. The data show that once Mehmed II established the official slave markets, this

trade restarted and prospered. It is therefore logical to say that the demand created by

Constantinople and later Istanbul was a significant factor in determining the intensity of

slave raids performed by the Crimean khans as it was driving force for demand.

Ottoman dependency on the revenue from taxing the slave trade should not be

underestimated. The Ottoman central budget counts taxes on the trade at 29 percent of

total Ottoman revenue in the years 1577-78.66 Within one century revenue from the six

major slave markets in the Empire amounted to almost one-third of the maliye’s67 budget.

Kołodziejczyk averages slave tax revenues from the ports of Akkerman and Očakiv alone

were based on a comparison between the exchange rate of ducats and akçe. The Porte’s Revenue for 1526-
27 was published in ducats, which I converted to akçe using a contemporary and relevant figure, in this
case the cost of Suleiman the Magnificent’s mosque which was 59 million akçe.
66
Inalcik, Social and Economic History of the Ottoman Empire, 283.
67
This was the Ottoman finance department in Istanbul, an office that employed a large number of clerical
offices and recorded Ottoman finances using an intricate double bookkeeping system. They had many
duties, the chief of which was to accurately record tax revenue and state expenditure. For more information
on the Maliye and its practices, see Linda T. Darling, Revenue-Raising and Legitimacy: Tax Collection and
Finance Administration in the Ottoman Empire, 1560-1660: Ottoman Empire and Its Heritage (Boston,
Brill Academic Publishers, 1996).

34
to be at least 100,000 akçe per year.68 These ports were significantly less prominent in the

region’s slave trade than were Uzak and Caffa and yet furnished large sections of the

overall Ottoman budget. These numbers reflect a startling reality of slave ownership in

the Ottoman Empire. A look at both elite and common strata reveal a culture inundated

with slaves and with an ever increasing demand for them.

Such a large source of revenue necessarily needed to be maintained and indeed

increased. This was not lost on the Porte. Throughout the sixteenth and seventeenth

centuries the sultan maneuvered to increase the Porte’s power in Crimean internal affairs

through both intervention and strong suggestion. The khan’s legitimacy may have been

based on a Ghengisid bloodline and approval by the Karachi begs, but ultimately the

financial backing of the sultan became necessary to the point that the sultan actually

practiced veto power in the process of succession in the khanate by the late sixteenth

century.69

Ottoman control over Crimean affairs was powerful but far from absolute, and it

eventually declined entirely after the period of study. From the late fifteenth century to

the late sixteenth century, however, the sultans’ ”liens de vassalit é”70 allowed them

influential backing in khanate succession. Alan Fisher recognizes the importance of the

slave trade to the Crimean Khanate but does not tie it to an increasingly economically

integrated relationship between the Porte and the khan. He asserts rather that the khans

68
“Bernard Pretwicz i jego apologia na sejmie 1550 r.” [Bernard Pretwicz and his apology in the diet of
1550], Biblioteka Warszawska, 1866, III, no. 7, p. 44-59, as quoted in Dariusz Kołodziejczyk, “Slave
Hunting and Slave Redemption as a Business Enterprise: The Northern Black Sea Region in the Sixteenth
to Seventeenth Centuries,” Oriente Moderno 25 (2006): 154.
69
Paul R. Magocsi, History of Ukraine: The Land and Its Peoples 2nd and Revised Edition (University of
Toronto Press, 2010), 15.
70
Bennigsen et al, Le Khanat de Crimée dans les archives du Musée du Palais de Topkapi, 5.

35
remained financially independent from the sultan throughout their existence. The idea of

a khanate with a central budgeting authority independent from direct Ottoman meddling

is very probable. However, between the Ottoman influence on khanate succession,

imperial revenue from Crimean trade, and the manner in which the sultan is addressed in

several pieces of correspondence, the evidence suggests suggests that at least before the

eighteenth century, the khanate deferred to the sultan’s authority on a number of matters.

Ottoman Islamic Tradition, Slavery, and Manumission

The Ottoman Empire was an Islamic empire. One of the eight tenets of the Ottoman

Circle of Equity71 states that the sultan was the symbol of justice in the empire, which

was legitimized by his adherence to the Islamic faith. The sultan was thus very careful to

at least be seen upholding Islamic principles. The Porte employed cohorts of religious

scholars to review and interpret new firmans.72

It is odd then to realize how widespread slavery was in Anatolia. While shari’a

law permits slavery, it is a very limited form of slavery and manumission of slaves is

considered pious, whether or not the emancipated party is Muslim. This is not an

apologist narrative. Emancipation was extremely common. Several fatwas were issued

in the course of Ottoman history dealing with manumission of slaves. One such law

stated that if a master frees a slave while drunk or in jest that the words could not be

71
Itzkowitz, Ottoman Empire and Islamic Tradition, 88.
72
Firman is a Persian loan word into Turkish that roughly means “demand” or “decree.” Firmans were the
sultan’s most important tool in administering the empire. These policies could range from Suleiman I’s
sumptuary laws to fiscal policy to justifications for wars or retreats.

36
taken back and that many slaves gained legal protection after such events.73 There is little

hard data to determine whether or not the volume of slave manumission in Anatolia

remained uniform across three centuries. The defterleri74 court inheritance registers held

by courts do not necessarily mention property that was lost, that is, a manumitted slave.

What the Kadi court records do mention are some common practices and

conditions for manumission. Slavery in the Ottoman Empire was rarely a life-long event

(although this certainly did happen) and the court records of Istanbul and Bursa are filled

with contracts that would seem to parallel indentured servitude. Ben Naeh’s study of

Jewish slave ownership in the Ottoman Empire emphasizes this point and discusses a

number of these contracts.75 Both Muslim and Jewish communities dealt with religious

restrictions on interactions with slaves. The Kadi records almost exclusively list female

slaves of Slavic origins.76 Just as Muslim prescriptions for the treatment of slaves

encouraged manumission as a pious act, Jewish slave owners were subject to several

rules regarding their female slaves, the most important of which was that the slave had to

be sold or manumitted once she had engaged in sexual intercourse with her owner.77

These laws and social attitudes favoring manumission are the reason why

Anatolia required a constant and growing supply of slaves. As slaves were married off to

freedmen or as their children became legally freed, demand for slaves increased. On the

state’s end of business we can trace a few events which fed this demand for slaves,
73
Catalcali Ali, Fetava, 1:60-2; Imber, “Hanafi Law of Manumission,” typescript, 2, and his “Involuntary
annulment of Marriage and Its Solution in Ottoman Law,” in Colin Imber, Studies in Ottoman History and
Law (Istanbul, 1996), 217-51. As quoted in Zilfi, Women and Slavery in the Late Ottoman Empire, 114.
74
Inheritance registers held by religious courts throughout all of Anatolia. Many of Turkey’s major cities
have well preserved archives of these registers and are excellent sources for further scholarly work,
especially in consumption studies.
75
Ben Naheh, “Blond, tall, with honey-colored eyes,” 325.
76
Ibid.
77
Ibid.

37
namely fundamental changes in agricultural and land settlement practice as the empire

expanded. Likewise, we can look at lulls in the slave trade and in our graph and compare

them to historical events which help us to explain them.

Changes in Agricultural and Price Realities

The Ottoman Empire was a highly centralized patrimonial system. The majority of

its revenue was exacted in the form of taxes on peasant labor based on the Çiftlik.78

From the beginning of the empire to the first capitulations to the European powers79,

agricultural industry was not profit driven. Peasants were expected to farm enough to

feed their family and to pay the predetermined tax. This tradition protected peasants

from the classical feudal exploitation experienced by their European counterparts, with

several laws including the inability of the state to move free peasants to new lands

arbitrarily. As the empire expanded in all directions, especially after the reign of

Suleiman, the state settled borderlands and engaged in profit-driven Çiftlik. As peasants

did have a modicum of protection, establishing these farms was most easily achieved

with slaves.

Mehmed II’s plan to revitalize Istanbul and make it the new capital caused the

creation of Khasslar lands occupied mostly by slaves in roughly 163 villages surrounding

78
Çiftlik were the most basic unit of Ottoman administrative and taxation practices. Halil Inalc ik defines it
as a household consisting of all labor- capable males and cattle. The number of these two things
approximated the overall productive capacity of the family farm and the local Timar or tax farmer would
adjust their taxation on the household accordingly.
79
Capitulations were a series of economic agreements between the Porte and several European powers
taken from a tradition going back to the Mameluke Sultanate. In general they decreed that foreign
merchants acting on behalf of participating kingdoms were not subject to regular Ottoman customs taxes
and had certain rights in Ottoman courts. For more on the capitulations see Halil Inalcik and Quataert, An
Economic and Social History of the Ottoman Empire: vol. 1, 1300-1600

38
Istanbul. These profit-driven farms caused some of the first major surges in demand for

slaves in Anatolia, which coincides directly with the beginning trend in the graph, the

conquest of Constantinople, and the establishment of the slave market in the years

following 1453.

Khasslar lands decreased in use as time wore on from Mehmed II and the

bureaucratization of the Crimean slave trade, so it was certainly not the only driving force

in Anatolia for slaves. The beginning of the sixteenth century saw the inception of

several popular revolts, namely the Celali Revolts.80 These revolts, which began in 1595,

were a response of overtaxed peasants and de-housed timar holders to currency

devaluation and the subsequent increase in taxes. The revolts resulted in intermittent

banditry throughout the sixteenth century and depopulated much of Anatolia’s farmland.

The revolts occur in the middle of the graph’s gap rather than the beginning. Therefore

the revolts did not initiate the downturn but remain a significant part of its continuation.

Coincidentally, these dates approximately describe the “dip” in figure 1.1. The

ensuing chaos of these popular uprisings in Turkey disrupted much of the economy from

1595 to 1610. This time period, during which commerce in major city centers in

Anatolia were disrupted by a series of Janissary and peasant sieges, highly damaged the

economy and the Porte became financially distracted in mitigating the effects of the

revolts while also trying to rebuild. This necessarily required a cut in spending, at least

briefly, on the purchase of slaves.

80
Omer Lütfi Barkan and Justin McCarthy, “The Price Revolution of the Sixteenth Century: A Turning
Point in the Economic History of the near East,” International Journal of Middle East Studies 6, No. 1
(Jan., 1975), 3-28.

39
 This unrest is the reason why the second graph is divided into two trend lines before

and after this period. While the graph shows that slave raiding as a variable dependent

upon Anatolian demand still increased, the revolts which occurred in the middle of this

period were a part of its temporary downturn. This dive is large enough to suggest an

overall negative trend when not taken with the outliers. Despite this difficulty in data,

over 65 percent of the points on the graph suggest a positive trend before and after this

traumatic event.

Popular consumption likely decreased immensely during this period due to both the

Celali revolts and what Barkan and McCarthy call the rise of European market

dominance. He utilizes price registers from Edirne to show that consumption in general

decreased as a result of heavily inflated prices. This period marks the start of a forty-six

percent average price inflation.81 While slaves are not specifically mentioned in the

account, silk and grains are. It is reasonable to assume that if bulk consumption of these

commodities diminished, then conspicuous consumption such as of slaves would also

decrease.

Slave ownership in the Ottoman Empire was by no means rare. It was, however,

subject to heavy regulation from religious and state authorities. Cultural attitudes toward

manumission and piety complemented these regulations, creating an environment that

could feasibly cycle out slaves every five years and require replenishment from non-

Muslim, outside sources, namely the services provided by the khanate.

81
Barkan and McCarthy, “The Price Revolution of the Sixteenth Century,” 9.

40
 As earlier noted, slave consumption was not necessarily uniform, although in

general it tended to increase. The low points of consumption are easily explained by

Anatolia’s greatest economic crisis since its rough beginnings. The Celali Revolts and

the rise of the European market system brought inflation, domestic disorder, and a

distinct fall in consumption as citizens could no longer afford conspicuous consumption

of slaves and the government was more concerned with maintaining stability in the

empire.

41
 CHAPTER IV

CONCLUSIONS

Over the course of three centuries the Crimean Khanate raided and pillaged Southern

Russia in order to feed an ever-growing slave market which was centered in Istanbul.

Both parties had specific interests which defined how the trade in this taboo commodity

would affect their states. As slave raiding became the major source of income for the

Crimean Tatars, the frequency of raiding became directly influenced by Anatolia’s need

for slaves. Consequently, events and attitudes in Anatolia are reflected in a graph of this

commodity. The sheer volume of slaves used in Anatolia in both elite and common

circles shows that slaves were in demand. Islamic attitudes toward slavery, especially the

practice of manumission, which created a non-recycling slave population that had to be

replaced from without, shows partially why this demand increased. The

bureaucratization of the slave trade by the Porte once they learned they could make

immense amounts of money by taxing the trade further emphasized the increase in

demand for slaves. More slave raiding and trading meant more tax revenues from the

Porte. Inversely, the one major downward fluctuation in the slave trade represented on

the graph happens to partially coincide with the Celali Revolts, which provided the

42
greatest amount of domestic unrest the Empire had seen since the incursions of

Tamerlane and the ten year interregnum period. Despite this dip, the data afterward still

trend upward, therefore falling in line with the theory that even through economic

instability, the demand for slavery and hence slave raiding trended upward.

The numbers presented in this study, both the estimations of slaves taken by the

Crimean Khanate per annum and the amount of money made by the Porte from this trade,

are the best estimations at present available to show a direct correlation between Ottoman

demand and financial power and the slave raiding activities of the Crimean Khanate.

This study lacks one very significant event which dates directly to the data. In

1571, the Crimean Khanate, led by Devlet Giray I, sacked Moscow and killed or captured

thousands.82 The data derived from Fisher confirms this number to be one hundred

thousand. This was a significant event and a scholarly study into this event would clarify

the Crimean-Ottoman relationship further and should be a continued topic of study in the

near future. I have chosen not to include this event because I wish to focus exclusively

on Crimean-Anatolian relations.

Ultimately the correlations of data between khanate slave raiding and Anatolian

economics suggest a very clear but perhaps unintentional relationship of mutual

economic interdependency. The more economically powerful partner of the two, the

Ottoman Empire, unconsciously affected the power structure and the economic activities,

namely slave raiding, of the other due to this interdependence. This relationship is

important to understand and know because it links events in disparate localities. It links

82
Fisher, The Crimean Tatars, (Hoover Institution Press, 1987), 45.

43
several economic circles, elite and popular. Most importantly, it suggests a core-

periphery relationship between the Ottoman core and the Crimean periphery.

44
 REFERENCES

Primary Sources

Bennigsen, Alexandre et al. Le Khanat de Crimée dans les archives du Musée du Palais de
Topkapı. Paris: Ecole des hautes etudes en sciences sociales, 1978.

Idem. L’empire ottoman et les pays roumains. Paris: Editions de l'Ecole des hautes études en
sciences sociales: 1987.

Berindei, Gilles & Minhea Veinstein, “Reglements de Suleyman Ier concernant le liva de Caffa,”
Cahiers du Monde russe et soviétique, 16, No. 1(1975): 57-104.

Dankoff, Robert & Kim Sooyong. An Ottoman Traveler, Selections from the Book of the Travels
of Evliya Celebi. London: Eland Publishing, 2010.

Fisher, Alan W. A Precarious Balance: Conflict, Trade, and Diplomacy on the Russian-
Ottoman Frontier. Istanbul, Turkey: The Isis Press, 1999.

Secondary Sources

Akarli, Engin Deniz. “Ottoman Historiography,” Middle East Studies Association Bulletin 30,
No. 1 1996, 33-36.

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

Barkan, Lutfi O. and Justin McCarthy. “The Price Revolution of the Sixteenth Century: A
Turning Point in the Economic History of the near East,” International Journal of Middle
East Studies 6, No. 1 (Jan., 1975), 3-28.

Barta, G ábor, Introduction and notes by G ábor Barta and Pál Fodor. Translated by Gábor Barta
and Jozsef Kun. Boston: Brill Publishing, 1996.

45
Cuno, Walz. Race and Slavery in the Middle East. New York, NY: The American University in
Cairo, 2010.

David, Geza. “Manumissioned Female Slaves at Galata and Istanbul around 1700,” Studies in
Ottoman Society and Culture (2002): 229-236.

David, Geza and Pal Fodor. Ransom Slavery along the Ottoman Borders. Boston: Brill
Publishing, 2007.

Davis, Robert C. Christian Slaves, Muslim Masters: White Slavery in the Mediterranean, the
Barbary Coast, and Italy, 1500-1800. Basingstoke, England: Palgrave Macmillan, 2003.

Darling, Linda T. Revenue-Raising and Legitimacy: Tax Collection and Finance Administration
in the Ottoman Empire, 1560-1660: Ottoman Empire and Its Heritage. Boston: Brill
Academic Publishers, 1996.

Faroqhi, Saraiya. ”Making a Living in the Ottoman Lands, 1480-1820,” Ottoman and Turkish
Studies 18 (1995).

Fisher, Alan W. The Crimean Tatars. Stanford: Hoover Institution Press, 1978.

Idem. “Slavery & Abolition: A Journal of Slave and Post-Slave Studies,” 13, no. 1 (1980).

Idem. The Sale of Slaves in the Ottoman Empire: Markets and Taxes on Slave Sales, Some
Preliminary Considerations, Beseri Bilimler-Humanities. Vol. 6, 1978, p. 171

Fleet, Kate. “Tax Farming in the Ottoman Empire,” The Medieval History Journal 6 (2006).

Inalcik, Halil. An Economic and Social History of the Ottoman Empire: vol. 1 1300-1600.
Cambridge: Cambridge University Press, 1997.

Idem. “The Struggle for the Eastern European Empire 1400-1700,” Turkish Yearbook of
International Relations 21 (1995).

Idem. Sources and Studies on the Ottoman Black Sea, Volume I: The Customs Register of Caffa,
1487-1490. Boston: Harvard Ukrainan Research Institute Publication, 1997.

Idem, “The Khan and the Tribal Aristocracy: The Crimean Khanate under Sahib Giray I,”
Harvard Ukrainian Studies ¾ (1979-1980): 452.

Itzkowitz, Norman. Ottoman Empire and Islamic Tradition. Chicago: University of Chicago
Press, 1980.

46
Karpov, S.P. “New Documents on the Relations Between the Latins and the Local Populations in
the Black Sea Area, 1392-1462,” Dumbarton Oaks Papers 49 (1995): 33.

Kizilov, Mikhail. “The Slave Trade in the Early Modern Crimea From the Perspective of
Muslim, Jewish, and Christian Sources,” Journal of Early Modern History 11 (2007): 1.

Kołodziejczyk, Dariusz. The Ottoman Survey Register of Podolia (ca. 1681). Boston: Harvard
University Press, 2004.

Idem. “Slave Hunting and Slave Redemption as a Business Enterprise: The Northern Black Sea
Region in the Sixteenth to Seventeenth Centuries,” Oriente Moderno 25 (2006): 149.

Kortepeter, C.M. “Āāzī Girāy II, Khan of the Crimea, and Ottoman Policy in Eastern Europe and
the Caucasus, 1588-94,” The Slavonic and East European Review 44, No. 102 (1966),
139-166.

Idem. “Ottoman Imperial Policy and the Economy of the Black Sea Region in the Sixteenth
Century,” Journal of the American Oriental Society 86 (1966): 87.

Magocsi, Paul R. A History of Ukraine: The Land and Its Peoples. 2nd Revised Edition.
Toronto: University of Toronto Press, 2010.

Manz, Beatrice F. “The Clans of the Crimean Khanate 1466-1532,” Harvard Ukrainian Studies 2
(1978): 286.

Matsuki, Ezio. “The Crimean Tatars and their Russian Slave Captives: An Aspect of Muscovite-
Crimean Relations in the 16th and 17th Centuries,” The Mediterranean Studies Group
Hitotsubashi University 18 (March 2006): 177.

Murray, Steven O. Islamic Homosexualities: Culture, History, and Literature. New York: New
York University Press, 1997.

Pamuk, Şevket. The Ottoman Monetary System and Frontier Territories in Europe, 1500-1700.
Ataturk Institute, Boazici University, 2000.

Rotman, Youval. Byzantine Slavery and the Mediterranean World, Cambridge, Mass. Boston:
Harvard University Press, 2009.

Seng, Yvonne J. “Fugitives and Factotums: Slaves in Early Sixteenth-Century Istanbul,” Journal
of the Economic and Social History of the Orient 39, No. 2 (1996), pp.136-169.

Toledano, Ehud R. As if Silent and Absent: Bonds of Enslavement in the Islamic Middle East.
New Haven, CT: Yale, 2009.

Idem, Slavery and Abolition in the Ottoman Middle East. Seattle, WA: University of
Washington Press, 1998.

47
Zilfi, Madeline C. Women and Slavery in the Late Ottoman Empire: The Design of Difference.
Boston: Cambridge University Press, 2010.

48
 APPENDICES

Fig 1.1

49
Fig 1.2

Fig 1.3

50
Fig 1.4

Analysis on Figure 1.1 by David Phillip Arthur Craig

To assess the correlation between time and the number of captives, a linear regression was
performed on these data. However, as number of captives is not a continuous quantity and as the
required normalityty and homogeneity assumptions are clearly violated by these data, a significance
assessment was not performed. A slight positive trend was observed when considering all data
points from Figure 1(r = 0.115; n = 36; R2 = .0132); however, this trend was driven
driv by a single,
outlying data point from 1676, for removing this data point produced a very slight negative trend
as depicted in Figure 2 (r = -0.0412;
0.0412; n = 35; R2 = .0017). Clearly, assessing the direction of the
observed trend largely depends on a single, outlying data point. The main question becomes,
based on evidence provided above, if a difference in number of captives prior to 1575 is observed
when compared to the number of captives after 1575. To begin to probe this question, two
regressions were performed
formed for data points occurring before or after 1575. Reducing the range of
data points predictably inflated the observed trends as depicted in Figure 3; prior to 1575, a lower
relationship between time and number of captives (r = 0.04349; n = 15; R2 = 0.1892)
0. was
observed compared to the relationship between time and number of captives after 1575 (r =
0.3787; n = 21; R2 = 0.1892). Again, this effect was driven by the outlying data point from 1676;
with this point removed, the relationship between time and number of captives was reduced when
considering data points after 1575 (r = 0.1674; n = 20; R2 = 0.0271) as depicted in Figure 4.
Simply stated, assessing differences between the number of captives before 1575 compared to
after 1575 is not conclusive when using an aggregate assessment (e.g. regression) due to the
outlier from 1676.

As some sort of distribution comparison is required in most investigations in the humanities, a

significance assessment is required. However, given the violations of continuity, normality, and

51
homogeneity assumptions required for null hypothesis significance testing, as well as the
pronounced, over-exaggerated effect of the outlying data point from 1676, a non-aggregated data
analysis technique is required to properly assess these data. Observation Oriented Modeling
(Grice, 2011; Grice, Barrett, Schlimgen, and Abramson, 2012) is a data analysis technique allows
comparisons of observed data to expected patterns of outcomes and then to evaluate the
differences with an accuracy index and a randomization test. This technique has been successfully
used in behavioral learning data (Craig, Grice, Varnon, Gibson, Sokolowski, and Abramson,
2012; Dinges et al., 2013). Observation Oriented Modeling (OOM) assesses individual
observations and does not rely on traditional summaries of data such as measures of central
tendency or variability. Using these methods successfully eschews the assumptions of null
hypothesis significance testing (e.g. continuity, normality, homogeneity).

Within OOM, two ordinal analyses were conducted. Ordinal analyses produce a percent correct
classification (PCC) value and a chance-value (a probability statistic). For each analysis, an
observed PCC value was computed by comparing an a priori ordinal prediction with direct pair-
wise comparisons of the observed data. The resulting PCC value ranges from 0 to 100 and is the
percent of the observed data that matches the expected pattern. Higher values indicate more
observations were correctly classified by the prediction. Next, a randomization process wherein
the observed data were randomly shuffled between groups/conditions was repeated 1000 times
for each ordinal analysis to create a range of randomized PCC values. The observed PCC values
were then compared to the randomized range of PCC values to compute a chance value (c-value).
The c-value ranges from 0 to 1 and displays the proportion of randomized versions of the
observed data that yielded PCC values greater than or equal to the observed data’s PCC value.
For example, a c-value of .001 indicates the observed PCC value was larger than 999 of the PCC
values obtained from 1000 randomized versions of the data. However, as c-values are calculated
from randomizations of the observed data points, each PCC value is assessed on an adaptable
distribution that is based on observed data rather than a hypothetical distribution (e.g., the
standard normal curve). Dinges et al. (2013) describes numerous philosophical and practical
differences between Observational Oriented Modeling versus null hypothesis significance testing
and contains and compares data sets analyzed via both methods.

To perform this assessment, the data were split at 1575 and two predictions were made. The first
prediction agrees with the provided a priori theory and posits more captives would be observed
before 1575 compared to after 1575. Indeed, this prediction was supported (PCC = 65.05; c-value
< 0.001). The second prediction counters the provided a priori theory and posits fewer captives
would be observed before 1575 compared to after 1575. Indeed, this prediction was not supported
(PCC = 33.33; c-value = 1.00). In short, 65.05% of the observed individual data points fit the
prediction that more captives were observed before 1575 compared to after 1575.

OOM Write-up for figure 1.1 by David Phillip Arthur Craig

52
Combination Observations created for Empty Data Set

Orderings :

ord_1

ord_2

Every observation (as it appears in the Data Window) in each of the 2 orderings (or cases) was
paired with every

other observation in the 2 orderings (or cases). This process created 441 new ordered
observations (cases).

{New data set of combinations created: See Data Window}

Ordinal Pattern Analysis for Combination Orderings

Classification Imprecision value = 0

Ordinal Classifications = Full pattern of ordinal relations

Missing Values = Omitted from Totals

Analyze: Entered Numbers Analyzed

1450-1600 Min = 2000.00 Max = 200000.00

1600-1750 Min = 1400.00 Max = 400000.00

53
Ordering Frequency Summaries

1450-1600 Units : 11 Missing : 126 Observations : 315

1600-1750 Units : 21 Missing : 0 Observations : 441

Totals Units : 32 Missing : 126 Observations : 756

Defined Pattern : All Observations

1450-1600

| 1600-1750

+ O Highest

O + Lowest

Classification Results : All Observations

Ordinal Relations between Pairs of Observations Classified According to the Defined

Pattern(s)

Classifiable Pairs of Observations : 315

Correct Classifications : 205

Percent Correct Classifications : 65.08

Classifiable Complete Cases : 315

54
 Correctly Classified Complete Cases : 205

Percent Correct Classified Cases : 65.08

Randomization Results : All Observations

Observed Percent Correct Classifications : 65.08

Number of Randomized Trials : 1000

Minimum Random Percent Correct : 40.00

Maximum Random Percent Correct : 59.37

Values >= Observed Percent Correct : 0

Model c-value : less than ( 1 / 1000); that is, < 0.001

Observed Percent Correct Classified Cases : 65.08

Number of Randomized Trials : 1000

Minimum Random Percent Correct Cases : 40.00

Maximum Random Percent Correct Cases : 59.37

Values >= Observed Percent Correct Cases : 0

Model c-value : less than ( 1 / 1000); that is, < 0.001

Ordinal Pattern Analysis for Combination Orderings

55
Classification Imprecision value = 0

Ordinal Classifications = Full pattern of ordinal relations

Missing Values = Omitted from Totals

Analyze: Entered Numbers Analyzed

1450-1600 Min = 2000.00 Max = 200000.00

1600-1750 Min = 1400.00 Max = 400000.00

Ordering Frequency Summaries

1450-1600 Units : 11 Missing : 126 Observations : 315

1600-1750 Units : 21 Missing : 0 Observations : 441

Totals Units : 32 Missing : 126 Observations : 756

Defined Pattern : All Observations

1450-1600

| 1600-1750

O + Highest

+ O Lowest

56
Classification Results : All Observations

Ordinal Relations between Pairs of Observations Classified According to the Defined

Pattern(s)

Classifiable Pairs of Observations : 315

Correct Classifications : 105

Percent Correct Classifications : 33.33

Classifiable Complete Cases : 315

Correctly Classified Complete Cases : 105

Percent Correct Classified Cases : 33.33

Randomization Results : All Observations

Observed Percent Correct Classifications : 33.33

Number of Randomized Trials : 1000

Minimum Random Percent Correct : 40.32

Maximum Random Percent Correct : 57.46

Values >= Observed Percent Correct : 1000

Model c-value : 1.00

Observed Percent Correct Classified Cases : 33.33

57
 Number of Randomized Trials : 1000

Minimum Random Percent Correct Cases : 40.32

Maximum Random Percent Correct Cases : 57.46

Values >= Observed Percent Correct Cases : 1000

Model c-value : 1.00

58
 VITA

Shawn Christian Broyles

Candidate for the Degree of

Master of Arts

Thesis: SLAVERY AND THE OTTOMAN-CRIMEAN KHANATE CONNECTION

Major Field: World History

Biographical:

Education:

Completed the requirements for the Master of Arts in history at Oklahoma State
University, Stillwater, Oklahoma in May, 2014.

Completed the requirements for the Bachelor of Arts in history at Oklahoma

State University, Stillwater, Oklahoma in 2010.

Experience:

Teaching Assistant at Oklahoma State University for two and a half years.

Taught upper division Islamic Civilization (HIST 3503) for one semester.