GREENBERG:Hausa Verse Prosody 135

localities. Following the playing of the verse in 2. ba mabanf ne ba saboda 'ayki ne

her praise, any young man may indicate his 3. h/ad/ariya taso
interest in her by putting a coin on the drum as 4. gada mugun nama
a gift to the performers. After this, the young
girl does a solo dance in front of the drum. The 1. An ax with which to begin work.
other occasion on which this drum is employed
2. It is not because of medicine; it is through
is cooperative farm labor, the gayyd. The kirari's
work.
of the various family heads of the community are
3. The thunderstqrmhas risen.
played to spur them on in their agricultural labor.
Both the quantity and tone of the spoken text is 4. The boar is bad to eat.
indicated very consistently in this type of drum-
ming so that all that will be necessary is the The material cited in this section is represen-
tative of original African types of poetic expres-
citation of the texts fully marked in regard to tone.
sion still surviving among the Hausa alongside
Since length of syllables is conveyed by the length
of the pause which follows the drum beat (no of patterns imported from the Islamic world. In
sustained tone, of course, is possible on a per-addition we might have cited the songs performed
to call spirits in the native possession rites known
cussion instrument) there is no way of indicating
the quantity of the last syllable of the phrase.as Bori. Although these performances have incor-
porated some Mohammedan elements, they are
kcirari'sof Young Girls essentially African in content. We have no record
1. kucciyar mayseka da 'alharmni of the music employed, but the songs are con-
2. da malamay take tafi 'a ta sa ta' wanka sidered to be kcirari'sof the respective spirits. The
3. siliki zaran kawna spoken forms of the text reveal no quantitative
4. musuilmin karfe 'zurfA pattern upon analysis.
1. The dove of Mecca with a silken nest, From the evidence adduced here, we may ten-
2. She goes with a Malam; she drank and washed tatively draw a contrast between the poetry of
herself (that is, she bought a medicine). Islamic origin which displays quantitative patterns
3. Silk is the thread which is desired, and in which the tone levels of the spoken text
4. Silver is the Muslim metal. are ignored in musical performance, and the
African poetry which does not exhibit quantitative
kirari's of Family Heads patterns and which incorporates both quantitative
1. gatarf masomin 'ayki and tonal distinctions into the musical settings.

THE CIRCASSIANS IN THE MAMLUK KINGDOM *

DAVID AYALON (NEUSTADT)

TbL-AVIV,ISRAXL

BARQfTQ,WHOMADE his fellow-Circassians the found in Egypt itself, or outside it, in the country
ruling caste in the Mamlfuk Kingdom, brought of origin of the Mamlflks?
about one of, if not the greatest racial transforma- Although Mamlfik sources deal at length with
tions ever witnessed in that state since its foun- the different Mamlfukraces and furnish extensive
dation. This transformation, as will be seen below, accounts concerning their countries of origin and
led to far-reaching changes in the organization of the Mongol Khans, they devote but little attention
the state. What were its causes? Were they to be to the internal situation in those countries. They
describe in detail the process of change in the
* This composition of the Mamluk army at the end of the
paper is part of a chapter entitled 'The
Mamliuk Races' from a work on the Mamlufik Army. fourteenth century; but they give the impression
 136 AYALON: The Circassians in the Mamluak Kingdom

that it was due solely to factors existing within more Mamlfiks each). Moreover, in the wake of
the boundaries of the Mamluk Kingdom, and not the males followed a considerable stream of women
to any outside agency. and female slaves belonging to the same races. It
A. N. Poliak, basing himself on Russian sources, should also be borne in mind that the emigration
considers that the decline of the Golden Horde was confined to one age-group, namely, adolescents,
during the latter half of the fourteenth century, so that the resultant harm to the population was
and the internal wars that broke out there, were the far greater than if the same number of emigrants
main causes of the transfer of the Mamluks' pur- had been distributed among all age-groups.
chasing-center to the Caucasus.1 As far as Arabic The Circassians are designated in Mamluk
sources are concerned, the present writer knows of sources as Jarkas or Jardakisa 3 (sing. Jarkasi).
only one reference to the destruction of the Kip- There are also alternative spellings: Carkcas or
chak Steppe, and that is by Timfir Lang's bi- Cardkisa 4 (sing. (arkasi) ; Sharkas or Sharakisa 5
ographer, Ibn 'Arabshah. This piece of information (sing. Sharkasi) ; 6 and less frequently al-Jihdrklas.7
is of the utmost importance, as it describes how Circassia is variously known as bildd al-Jarkas,8
the Kipchak, a flourishing territory enjoying bildd Jarkas,9 or simply Jarkas; 10 and occasionally
security and providing ample room and grazing- as jabal al-Jarkas.11 As regards racial origin, the
grounds for a very varied nomad population, was Circassians are said, by one historian, to be min
turned into a wilderness and its inhabitants greatly al-Turk ('of the Turks'), and 'they are a well-
reduced in number as a consequence of the wars known tribe to the genealogists.' 12 The tribe of
between Tuqtamish and Aydaka and the campaign Jarkas is described as dwelling in bildd al-Shimdl
of Timur Lang.2 At any rate, even if the suprema-
'in the hills surrounding the plain of the Kipcha-
cy of the Circassian race was due in the first place kis and the Rus and Alldns, on the eastern (?) side
to the will of the rulers of the Mamluk Kingdom,
the decline of the Kipchak population as a con- overlooking their plains.' 13 They live in poverty,
sequence of the above-mentioned wars contributed
s Ibn Taghri Birdl, hawddith al-duhar, ed. Popper,
appreciably to the preservation of that supremacy
down to the very end of Mamluk rule. It can also Berkeley, 1930, p. 699, 1. 14; Ibn Iyas, badd'i' al-zuhur
(vols. I, II, Cairo, 1311-1312A. H., vols. III, IV, V, Con-
be assumed that the decline in the Kipchak popu- stantinople, 1931-1932), vol. III, p. 2, 1. 5; ibid., vol. IV,
lation was due in no small measure to the constant p. 200, 1. 13. For sing. Jarkasi see al-Sakhawi, al-tibr
stream of emigration of young Mamluks on a scale al-masbuk, Cairo, 1896, p. 110, 1. 12.
4 Ibn Iyas, II,
sufficient to maintain for centuries the army of a p. 10, 1. 15. For sing. uaarkasisee tibr,
p. 110, 1. 2; Ibn Iyas, II, p. 48, 1. 13.
mighty Power (i. e. the Mamluks of the sultans 5 Ibn Qadi Shuhba, fol. 40B, 11. 18-19.
and amirs: some of the latter owned hundreds or 6 al-Sakhawi, al-daw' al-ldmi', Cairo, 1353-1355 A. H.,
III, p. 250, 1. 14; Ibn al-Furat, ta'rikh al-duwal wal-
1Revue des .tudes islamiques, 1935, pp. 241-242; Bul- muluk, ed. Zurayq, Beirut, 1936-1942, IX, p. 101, 11. 1-5.
7 subh, IV,
letin of the School of Oriental Studies, London, vol. X, pp. p. 429, 1. 6; ibid. VII, p. 293, 1. 2.
864-867. There is so far no adequate proof that the 8subh, IV, p. 416; ibid., p. 462, 11. 2-6; Ibn IHajar
Black Death caused a permanent decline in the Kipchaki al-'Asqalani, al-durar al-kdmina, Haydarabad, 1348-1350
population. A. H., III, p. 269, 1. 1, 1. 3.
9 Ibn Iyas, IV,
2 akhbdr timur, p. 113, 1.
5-p. 115, 1. 3; ibid., p. 126, p. 184, 11. 28-29; ibid., III, p. 145, 1.
1. 2-p. 127, 1. 4 (in the translation of J. H. Sanders, 10; daw', II, p. 270, 1. 1; ibid., VI, p. 201, 1. 1, p. 218,
Tamerlane, London, 1936, pp. 76, 77, 78). For the de- 1. 13, p. 221, 1. 3; ibid., X, p. 270, 1. 2, 1. 26.
struction by Timar of Sarai and other towns in the 10daw', III, p. 12, 1. 48, p. 61, 1. 1, p. 63, 1. 8, p. 284,
Kipchak, see ibid., p. 122, 1. 2. For large-scale move- 1. 18; ibid., VI, p. 229, 1. 9; ibid., X, p. 280, 1. 15.
ments of tribes carried out by Aydakf in order to mis- 11akhbdr timur, 273, 1. 2.
p.
lead Timar, see ibid., pp. 122-125. These enforced 12 Ibn Khaldfin: kitdb al-'ibar, Cairo, 1284, A. H., V,
movements doubtless also contributed to the general p. 472, 11. 10-11. The term Turk has sometimes very
chaos. On the flourishing condition and dense population elastic and loose meaning in contemporary sources. On
of the Kipchak during the 13th century and the first the two ordinary meanings of this term see note 19
half of the 14th century, see: al-Mufaddal b. abi al- below. Classifying the Circassians as Turks by race is
Fada'il, al-nahj al-sadid, in Patrologia Orientalis, vol. most unusual in Mamlfk sources. Usually they are
XII, pp. 457, 1. 4-458, 1. 2; al-Qalqashandi, subh al-a'shd, mentioned as different from and antagonistic to the
Cairo, 1353-1355 A. H., vol. IV, p. 451, 1. 11; Ibn Taghri Turk. The Circassians may be called Turk only as far
Birdi: al-manhal al-safi, Paris (de Slane No. 2068-2072), as this term is synonymous to Mamluk (see below).
13 Ibn
and Cairo MSS., vol. I, fol. 167A, 11. 10-12. Khaldun, V, p. 472, 11. 5-8.
 AYALON: The Circassians in the Maml2lckKingdom 137

and most of them are Christians.14 We find no Jarkas yd lahu nasab mudh ild Ghassdn yantasib
evidence in Mamluiksources of the military prowess 'How splendid a genealogy have the progeny of
of the Circassians, as we do with reference to the Jarkas, since they are related to Ghassan.'18 Apart,
ruling race in the early MamliukPeriod.15 however, from the two references quoted, the pre-
An unexplained tradition concerning the Cir- sent writer has not discovered any other allusions
cassians is to be found in Ibn Khaldun, according in Mamlfik sources to the tradition attributing the
to which they are descendants of the Ghassanis who origin of the Circasians to the banu Ghassdn.
entered bildd al-Ruim with Jabala b. al-Ayham at The Circassians were not newcomers to the
the time of Heraclius' retreat from Syria. Ibn MamluikKingdom when they assumed supremacy
Khaldun, however, considers the tradition to be at the end of the fourteenth century. There is
inexact, and gives what in his belief is the true intermittent mention of them in the Bahri period
version, namely, that the Circassians stem from the as being one of the more prominent racial groups.
Turk, and that they went to Circassia before the What is more important, they are described as a
arrival of the Ghassamnsin bildd al-Rutm. After the growing force of which the Turk 19-the members
death of Heraclius, the Ghassanis despaired of of the dominant race at the time-were afraid.
returning to their homeland. Internal strife broke No such statement is made with regard to any
out within the Byzantine Empire, and the Ghas- other Mamluikrace. The fact that the Circassians
sanis, seeking an ally, concluded a pact with the finally ousted the Turk and gained supremacy
Circassians and ' settled in the eastern plain of the shows that there was good reason for the fears
Circassian hills beyond Constantinople.' They en- that had been felt, and that the Circassians' desire
tered into relations of genealogy and marriage with for power had existed long before its successful
them, until the Ghassani tribes disintegrated and realization.
' went up to the Circassians in the hills and dwelt Mention is made of the Circassians as early as
with them. Nor does this seem unlikely, since a the end of the Ayyfubidperiod, when members of
number of Circassians acquired the nisbas of al- that race are recorded as being among the slaves
Ghassani.' 16 Thus, according to Ibn Khaldfin, purchased by al-Salih Ayyfib, founder of the
Barqfiq's father was called Anas (correctly, Anas) Bahrnya.20This fact, which appears in Ibn Khal-
al-Ghassni.17 That this tradition was known to duin, is not, however, mentioned by any other
others besides Ibn Khaldiun is shown by the verse historian. By the time of the founding of the
lamenting the destruction of the Mamliuk King- Burjiya 21 the Circassians had become a very promi-
dom at the hands of the Ottoman Turks: nasl nent element in this new corps 22 which occupied a

18 Ibn
14 subh, IV, Iyas, V, p. 193, 1. 3.
p. 462, 1. 5; Poliak, REI, 1935, p. 242. 19 The word Turk had two ordinary meanings in Mam-
15
See note 19 below. Poliak's etymology of jarkas is lfik times: (a) it was a synonym for the Mamlfiks, so
quite untenable, REI, 1935, Note 5; BSOS, vol. X, pp. that dawlat al-turk or dawlat al-atrdk meant the Mam-
865-866; Khazariya (in Hebrew), Jerusalem, 1943, p. 35. luk Kingdom; and (b) it was used as a generic name for
In his dictum, jarkas wama'ndhu bil-a'jamfya arba'at the whole of that element coming from the Kipchak
anfus (manhal, II, fol. 172B, 11. 10-11), Ibn Taghri Birdl plain which predominated in the Mamlfik Kingdom dur-
no doubt meant to say 'four persons,' since kas in
ing its earlier period, erroneously styled by orientalists
Persian means 'person.' At any rate, it is wrong to 'the period of the Bahri Mamlfiks.' On the military
translate the phrase as 'having four souls.' Minorsky prowess of the Kipchakis see subh, IV, pp. 457-58.
(BSOS, vol. X, p. 867) has already commented on this. 20
Ibn Khaldin, V, p. 373, 1. 8.
Whatever reading is given, Ibn Taghri Birdi's explana- 21 The Burjiya was a Mamlfik army
tion has nothing to do with 'courageous.' Thus we have corps founded by
al-Mansur Qalaun, and at the time of its formation it
to reject Poliak's assertion in his reply to Minorsky numbered 3,700 men. It was given that name because it
(Khazariya, p. 319, Note 15) that the explanation pub- was garrisoned in the towers (abraj) of the Cairo citadel.
lished by him 'is the only one substantiated by the In many respects the burjiya were the successors of the
sources.' Ibn Taghri Birdi, incidentally, was not the Bahriya, the corps d'elite formed by al-Malik al-Salih
first to provide this explanation of jarkas. He was Najm al-Din Ayyufb. The Circassian period is errone-
preceded by Ibn Khallikan (I, p. 213), from whom Ibn ously styled by orientalists 'the period of the Burji
Kathir (XIII, p. 63) copied it. Mamluks.'
6 Ibn Khaldfin, V, p. 472, 11. 5-18. 22 The composition of the Mamlufk army in
17
Ibn Khaldun, V, p. 472, 1. 4. The epithet al-Ghas- Egypt was
as follows: (i) The royal Mamlfiks (mamdalk sultdniya),
sani for Anas, father of Barqfiq, is unknown to us from who were subdivided into (a) the Mamluiks of former
Mamlfik sources. sultans (mam,alik al-salCttin al-mutaqaddima, qardnis or
138 AYALON:The Circassians in the Mamluk Kingdom

position similar to that formerly occupied by the silence concerning the Circassians. Suddenly, how-
Bahriya.23 One of the Burjl Circassians, Baybars ever, the following story appears: 'Ramaddn, 748
al-Jashnakir, even gained the sultanate. From a A. H. (during the sultanate of Hasan). Then the
number of casual remarks in contemporarysources, amirs decided to reduce the royal and other ex-
it can be seen that the Circassians were already a penses. Lists were drawn up of the state expendi-
thorn in the side of the ruling race. Al-Hajj ture, and the amirs began to sell those of the royal
Bahadur is reported as stirring up incitement Mamlfuks who were Circassians. It was Sultan
against Aqfsh al-Afram, saying to those who sided IHajji who showed them favour, at the instance of
with him: 'These Circassians! When they are Ghurlfi. He brought them from all quarters and
stronger than we, they will take our souls. Let us wanted to give them precedence over the Atrdk,
do the thing before it is done to us.' 24 When al- until they became very prominent among the amirs
Muzaffar Baybars came to the throne, al-Afram with their large turbans and influential positions.
was pleased. But al-Hajj Bahadur and other lead- They made themselves unusually large caps. All the
ing amirs were disgruntled and said: 'When the Circassians were seized and exiled in a degrading
Circassians grow powerful, they will destroy the manner, and people said: " These be the ones with
Muslims and the land' (yuAlikaf al-ibdd w-al- the hungry souls who stir up much discord. "' 29
bildd).25 The hostile attitude of the ruling Turk The purge does not, however, seem to have been
is also illustrated by the following. When al- very thorough, for by Shawwal of the same year
Afram, the Governor of Damascus, read the letter we hear of an exchange of correspondencebetween
informing him of the appointment of Baybars as the Circassian Mamluiksand amir Husayn, son of
sultan, he almost went mad with joy, as he had al-Nasir Muhammad b. Qalaun, offering to make
been his khushddsh26 and was also a Circassian, him sultan. Forty of the conspirators were caught
both being regarded at the time as alien to the and sent separately on swift camels to al-bilad al-
Atrak.27 There are several other allusions to this Shamiya. Later, another group was caught. These
feeling of strangeness.28 received a sound drubbing opposite the Iwdn, after
The gradual decline of the Burjiya is marked by which they were handcuffed and thrown into the
qardnisa); (b) the Mamlfiks of the reigning sultan
prison of Khizdnat Shama'il. From then on, as a
(mushtarawdt, julbdn, ajldb). From among the mush- precautionary measure, the only persons allowed to
tarawdt a corps of pages and bodyguards was selected enter the sultan's palace were amirs of the Supreme
known as khdssikiya; (c) the Mamlfiks of the amirs Council (umard' al-mashdra).30 In other words,
who passed into the service of the sultan owing to the this was an attempt on the part of the Circassians
death or dismissal of their masters (sayfiya). (ii) The
Mamlfiks of the amirs (mamdalk al-umard', ajndd al-
to effect a military coup d'etat some 35 to 40 years
umard'). (iii) The sons of the amirs (awlad al-nds) before Barquiq'ssuccessful attempt.
and soldiers drawn from among the local inhabitants Barqiuq, as we have seen, was the sultan who
(ajnid al-halqa). The Mamlfik amirs were divided into effected the greatest and most far-reaching racial
three classes: (i) amirs of ten; (ii) amirs of forty; (iii) transformation in the entire history of the Mamliuk
amirs of a hundred.
23 Two chapters of my work mentioned above are de- Kingdom. He was able to oust the Turk and
voted to the history and structure of the Bahriya and replace them with his fellow-Circassians.31 This
Burjiya. change was brought about without any major up-
24 manhal, II, fol. 96B, 11. 2-6.
heaval. Only during the initial period following
25durar, I, p. 500, 11. 11-14. the rise of the Circassians are there signs of racial
26 The khushddshiya were the Mamltuks bought and
liberated by the same master. The feeling of comrade- warfare against them, as we shall soon see. Their
ship among these Mamluks, and their loyalty to their sanguinary war against the Sultan Faraj did not
lord, who had bought and freed them, were one of the have a racial basis: it was directed against Faraj
chief foundations of the entire Mamlfik regime. This personally, as they objected to the appointment of
aspect is dealt with at length elsewhere in our work on one of the awldd al-nds as ruler over them. Their
the Mamlfik army.
27 Ibn
Taghri Birdi, al-nujam al-zdhira, Cairo ed.,
1928-1944, vol. VIII, p. 236, 11. 3-5.
29 nujum, V, p. 56, 11. 14-20.
28 manhal, I, fol. 206B, 11. 18-21; ibid., II, fol. 2A, 11.
30nujum, V, 57, 11. 22-p. 58, 1. 12.
p.
12-15; nujum (Cairo), VIII, p. 232, cf. p. 227, 11. 1-6,
31
subh, IV, p. 458, 11. 16-19; al-Suyufii, husn al-
muzhddara, Cairo, 1219 A. H., II, p. 105, 11. 6-8; nujum,
p. 233, 1.2; durar, I, p. 396, 1. 18; ibid., III, p. 246, 11.
5-6, p. 247, 11. 6-7. V, p. 362.
 AYALON: The Circassians in the Mamltlu Kingdom 139

success was so decisive that throughout the period The following is a description of the rise and
of their ascendancy the son of a Circassian sultan establishment of the Circassians in the Mamluk
did not occupy the throne for more than a very Kingdom.
brief period (usually a matter of months).32 Barqfiq was called al-qd'im bi-dawlat al-Jar&-
kisa.33 He bought Mamluks on an extensive scale,
32 The sons of the sultans, who occupied a special and preferred the Circassians to other races.34
position among the awldd al-nds (nujizm, V, pp. 159- From this it appears that he did not altogether
160; ibid., V, p. 293, Notes; hawddith, p. 142, 11. 2-3)
were called al-asydd, awldd al-mulilk, or al-asydd awldd refrain from purchasing Mamlfiks of other races,
al-mulak. Their style of address was sayyidi. The reign- -a fact which is corroboratedin several ways. In
ing sultans, especially in the Circassian period, dealt the first place, Turk as well as Circassians are men-
most rigorously with the sons of the former sultans, to tioned as fighting for him in 791 A. H.35 Secondly,
prevent them being used as tools by would-be usurpers. each sultan automatically took over his predeces-
Up till the time of Barsbay, most of the asyad were
confined to the Cairo citadel, and many of them had sor's kuttdbfya; 36 and as inheritor of Sultan
never seen Cairo and had no idea what a city looked .Hjji's or Sha'bdn's kuttdbiya Barqfiq had to main-
like. About 825 A. H., Barsbay ordered them to leave the tain several hundred Turk. Thirdly, we find in
citadel and take up their abode in the city. The bustling the biographical compilations and necrologies con-
life of Cairo, however, brought about their rapid de-
generation. Many of them became impoverished, and
tained in the chronicles of the Circassian period a
' all their glory departed' (manhal, I, fol. 55A, 11.
2-13; surprisingly large number of amirs belonging to
ibid., fol. 179A, 1. 21-fol. 179B, 1. 2; Ibn Iyas, II, p. 15, Barquiq who were of Turkish origin (turki al-
11. 11-14; daw', III, p. 53 11. 8-10). The more prominent jins)37 The number of amirs of Turkish origin
and dangerous among them, especially those who had
succeeded their fathers to the sultanate, were generally explanation for it. In one place he remarks: 'We have
sent by the Circassian Mamlfks to the prison at Alex- seen the same reward meted out time and time again,
andria (Ibn lyas, II, p. 14, 11. 25-26; and numerous other from the day that Barqiiq deposed al-Mansfir .Hajji down
authorities). Occasionally one of the Circassian sultans, to our own day. They all drink from the same cup
to demonstrate his large-heartedness, would have one of handed to them by their atdbeks, and the contents of
the sultans' sons brought from Alexandria to Cairo, the cup are prepared by their fathers' Mamluks. We
where he would have splendid entertainments held in his have mentioned this matter in many places, and it were
honor and even allow him to go on the pilgrimage to better to maintain silence concerning it' (nufjum, VII,
Mecca (hawddith, p. 706, 1. 10-p. 707, 1. 15; Ibn Iyas,
p. 419, 11. 2-6). Elsewhere he asserts his inability to
III, p. 66, 11. 12-16, p. 150, 11. 6-14, p. 152 11. understand why the sultan at the last moment designates
22-24). These however, were isolated instances, and his son to succeed him, since the sultan knows full well
imposed no obligation whatever on the sultans. The that his son will undergo the same fate as overtook the
position of the asydd remained as described above down son of the previous ruler (nujum, VII, p. 394, 11. 9-13;
to the end of the Mamlfik period. (For material on the see also nujum, V, pp. 228-230; ibid., VII,
asydd and their status in the Mamlfik Kingdom, see: pp. 394-396;
.hawddith, p. 134, 11. 1-2).
al-Maqrizi, kitdb al-sulak lima'rifat duwal al-mulak, 33nujam, V,
p. 362, 11. 2-7, p. 596, 1. 22.
Cairo, 1934-1942, II, p. 490, 1. 13; nujam, V, p. 228, 84 manhal, II,
p. 73A, 11. 19-20; nujum, V. p. 598,
11. 18-19, p. 229, 1. 3, p. 282, 1. 23, p. 320, 1. 21, p. 397, 11. 7-8.
11. 15-17, p. 505, 1. 12; ibid., VI, p. 266, 11. 2-3, 85See Notes
p. 432, 42, 45, 47.
11. 5-7, p. 514, 11. 7-8, p. 545, 11. 8-9, p. 772; ibid., VII, 86 The kuttdbiya were the
young Mamlfiks undergoing
p. 320, 1. 1, pp. 508-509, p. 511, p. 644, 11. 1-8, pp. 664- their training in the military school who had not
665, p. 678, 11. 8-9; hawddith, p. 149, 1. 1, p. 305; Ibn yet
been liberated (they were only freed on their
al-Furat, IX, p. 56, 11. 15-18, p. 91, 11. 18-21, p. 176, 1. 21; passing-
out (kharj) of the school). The appellation is
Ibn Iyas, II, p. 15, 11. 1-2, p. 60, 1. 4, p. 79, p. 108, 1. 1, appar-
ently derived from the word kuttdb, denoting a Moslem
p. 113, 11. 6-10; ibid., III, p. 188, 11. 10-12, p. 195, 1. 12; boys-school. Students of the Mamlfk period generally
ibid., IV, p. 9, 1. 7, p. 399, 11. 15-23, p. 406, 1. 9; daw', read the word as kitdbiya; but the suggested reading of
III, p. 53, 11. 8-10, p. 87, 11. 6-7, p. 201, p. 217; ibid., VI, kuttdbiya, given here, is based upon an explicit vocali-
p. 73; ibid., VII, p. 274; Khalil b. Shahin al-Zahirl, sation in an MS. (see: manhal, I, fol. 167A, 1. 17; ibid.,
zubdat kashf al-mamdlik, ed. Ross, Paris, 1894, II, fol. 53B, 1. 7; ibid., III, fol. 153A, 1. 16; zubda, p.
p. 111,
11. 5-12; subh, XIII, p. 167, 11. 16-19.) A surprising
116, 1. 12, p. 125, 1. 12. For the argument in favour of
feature about the Circassian sultanates is that even the the reading kitdbiya, see al-Maqrizi, al-mawd'iz wal-
later rulers, who were well aware of the fate that lay i'tibdr fi dhikr al-khitat wal-dthdr, Cairo, 1270 A. H.
in store for the sons of sultans appointed by their II, p. 213, 1. 33; Dozy, Supplement, under kitdba and
fathers to succeed them, did not draw the obvious con- kitabi).
clusion, and continued the practice in the certain knowl- a7Some
examples: nujum, VI, p. 298, 11. 9-11, p. 449,
edge that their sons would be deposed. This fact arouses 1. 17-p. 450, 1. 1; ibid., VII, p. 265, 11. 20-21, p.
the astonishment of Ibn Taghri Birdi, who can find no 639,
11. 6-7; manhal, II, fol. 156B, 1. 6, fol. 156A, 1. 22;
ibid.,
140 AYALON: The Circassians in the Mamluk Kingdom

under Barqfuq's son, Faraj, who was inimical to the rank who was not worthy of it.' 41
Is it to be
Circassians, is found to be far smaller.38 Never- inferred from this passage that the Kipchakis were
theless, the complete establishment of Circassian less insistent on racial purity than the Circassians?
domination was fairly rapid. As early a writer as The great variety of races in the Bahri period does
Qalqashqandi, who completed his book in 1412, indeed tend to confirm this view.
states: 'In our time most of the amirs and army The process of the establishment of Circassian
have become Circassians . . . The Turk Mamliuks supremacy exhibits a number of fluctuations and
of Egypt have become so few in number that all retrogressions, the first of which took place after
that is left of them are a few survivors and their Barqfiq's downfall and removal from the throne in
children.' 39 791 A. H. Circassian power, which had been built
The change in the racial composition of the up by Barquiqwith so much effort, suffered a heavy
Mamliuks effected by Barqfiq, and the consequences check; and but for his rapid return to the throne
it brought in its train, gave rise to strong criticism the establishment of complete Circassian domina-
on the part of the Mamliuk writers, who through- tion might have been considerably delayed.
out the Circassian period retained their liking for Although the wars of 791 A. H. are described
the Turk,40 and did not curb their propensity to fully in Mamlfik sources, one factor is almost com-
vilify the Circassians. Referring to Tashtamur pletely slurred over, namely, that the war pro-
al-'Alda, formerly dawadar and later atdbak al- claimed by Mintash and Yalbugha was not only
'asdkir (commander-in-chief), who was removed by aimed against Barqiuq himself, but also against his
Barka and Barqfiq, Ibn Taghri Birdi says: 'The Circassians. This can be learned principally from
time of Tashtamur was a flourishing and plentiful Ibn al-Furat, who was a contemporary of the events
time for the Mamliuk Kingdom under his wise he describes, and who gives a most detailed ac-
direction, and that condition prevailed until he count of the fighting of 791 A. H. Several impor-
was removed from office and thrown into prison. tant particulars in this connection can also be
In his place came Barqiuq and Barka, who did obtained from Ibn Qadi Shuhba. Although the
things in the kingdom from which the population war did not have the extermination of the Cir-
suffers to this day. Then Barqfiq became sole ruler, cassians as its object, it was at times waged with
and turned the affairs of the realm upside-down, great determination and vindictiveness.
and his successors have maintained his policy down After the defeat of Barqfiq's army in his first
to the present. For he gave precedence to the mem- encounter with his opponents in Syria, Yalbugha's
bers of his own race over the others, and gave those men, on coming across any Circassian Mamlfuk,
of his own Mamlfiks (ajldb) who were related to stripped him of his clothes, took his weapons and
him large fiefs and high offices while they were still accoutrements, and cast him into prison. At the
in their minority. This is the main cause of the same time any of Barquiq's Turkish Mamliuks who
decline of the kingdom. Indeed, is there anything fell into their hands they released, after despoiling
more grave than to set the minor over the senior? them of all their belongings.42 The despoliation
This is at variance with the practice of the former and incarceration of Barquq's Mamliuks continued
sultans; for they did not recognize the superiority after the fighting was over.43 During the fighting
of any one race. Whenever they found a man who in Cairo, amir Tuqtay, a eunuch, who fought with
displayed wisdom and courage, they showed him great gallantry on the side of Yalbugha and Min-
preference and favour. No one was given office or tash, cried out: 'Where are the Circassians with
their genital organs? Behold me, Tuqtay, the
III, fol. 8A, 11. 11-13; fol. 169A, 11. 17-22; fol. 177A,
41 manhal,
11. 12-18; ibid., V, fol. 43B, 1. 19, fol. 46B, 1. 1. For III, fol. 185B, 11. 14-23 (this passage has
the expression turki al-jins in the Circassian period, see been quoted by Poliak, REI, 1935, p. 242 Note 1). Of
nujum, VI, p. 547, 11. 13-14, p. 675, 1. 14, p. 803, 1. 15, Qalaun, the historian says that he did not evince leanings
p. 840, 1. 12, p. 853, 11. 5-7; ibid., VII, p. 309. towards a particular race, but selected able men, irre-
38
nujum., VII, p. 309, 1. 4, p. 564, 11. 8-12; manhal, spective of the races they belonged to. That is why his
II, fol. 18B, 1. 8; daw', II, p. 269, 1. 29. reign was so protracted, despite the diverse racial origin
39 subh, IV, of his Mamlfiks (nujum, Cairo, VII, p. 327, 11. 13-15).
p. 458, 11. 16-19; cf. ibid., V, p. 416, 11.
8-13. 42 Ibn al-Furat, IX, p. 64.
40 See authorities
quoted below.
43
Ibn al-Furat, IX, p. 64, 1. 24-p. 65, 1. 8.
 AYALON: The Circassians in the MamllukKingdom 141

eunuch! ' 44 Barqfiq's Turkish Mamluks joined the the purchase of Circassians on an increased scale,
enemy during the fighting in Cairo, 'and whoso- but regretted his action towards the end of his life,
ever was a Circassian was taken.' 45 Towards the when an attempt was made to assassinate him led
end even the Circassians began to desert Barquiq.46 by 'All Bay.55 This took place in dhu al-Qa'da,
Later, Yalbugha gave orders to seize all Barquiq's 800 A. H. 'From that day on relations between the
Mamliuks, whether Turks or Circassians.47 When sultan and the Circassian Mamlufkswere impaired.
Cairo was stormed, an assurance of safety was given He went into Ird ( ?), his favourite wife, a turktyat
to the Circassians if they came out of their hiding- al-jins, who warned him against buying Circassian
places and surrendered.48 Despite this assurance, Mamlfiks. Her advice to him was as follows:
however, Yalbugha exiled some of the Circassians " Make your army of varied pattern (ij'al 'askarak
who gave themselves up.49 ablaq), and let it be composed of four races, Tatar,
The anti-Circassian policy which marked the Jarkas, Rum and Turkmran,and then you and your
beginning of the war against Barqfuq was post- progeny will have respite." He replied: "Your
poned for a while from the moment the fight broke advice is sound, but this is what fate has decreed.
out between Mintash and Yalbugha to secure the Let us beseech the Almighty to mend this matter
sultanate. It now became of vital importance for from now on." ' 56
each of the contestants to attract as many suppor- Evidently Barqiuqwas unable to accomplish his
ters as possible. The Circassians went over to contemplated alteration of the racial composition
Mintash.50 As soon, however, as Mintash emerged of his Mamliuks,as he met his end a few months
victorious, he reverted to his original policy towards after the conversation recorded above. During the
the Circassians,-a course he was encouraged to reign of his son Faraj, who succeeded him, the
adopt by the signs that Barqufqwas preparing to Circassians passed through a very difficult crisis,
stage a come-back. He therefore sought to outwit and large numbers of them were killed off. The
the Circassians who had fought on his side in the underlying cause of the conflict between Faraj and
battle against Yalbugha. He invited them, 200 in his father's Circassian Mamluks was their desire
number, to the citadel on the pretext of distribut- to depose him and place one of their own number
ing pay among them. The moment they were in- on the throne. Attempts to do this were suppressed
side, he orderedthe gates to be closed and had them in rivers of blood.57 In one day alone Faraj
thrown into prison.51 He also announced rewards slaughtered over 100 Circassians and cast their
for those handing over Circassians who had gone bodies over the citadel parapet into the moat be-
into hiding.52 He then issued an order to the wdal low.58 On another occasion Taghri Birdi, the Gov-
to seize the Circassian Mamluks whereverthey were ernor, charged Faraj with killing 500 men, and
to be found. The wali succeeded in catching some told him that his very horse was rebelling under
of them.53 Before Mintash left Damascus, he killed him.59 Faraj's reply was that what had been done
Barquiq's Circassian Mamliuks imprisoned in the could not be undone, and that he had acted as he
citadel.54 had in order to protect his honor.60 According to
On his return to the sultanate, Barqfiq resumed Maqrizi, during 814 A. H. alone Faraj killed 630
Circassians.6' This historian goes on to say that
44
manhal, fol. 193A, 11. 14-21. Faraj brought about the destruction of the Cir-
45 Ibn Qadi
Shuhba, fol. 35A, 1. 12. cassians, and points out that the numerical value
46 Ibn al-Furat, IX, p. 88, 11. 2-6. of the letters forming the names Faraj and Jarkas
47 Ibn al-Furat,
IX, p. 93, 11. 6-15. is equal, so that when one is subtracted from the
48
Ibn al-Furat, IX, p. 95, 11. 24-25, p. 96, 11. 9-11.
49 Ibn al-Furat, IX, p. 101, 11. 4-5; cf. ibid., p. 129; 66 nujuim, V, p. 598, 11. 7-9.
nujum, V, p. 487, 11. 9-11; Ibn Qadi Shuhba, fol. 35A, 56 nujim, V, p. 585, 11. 3-8.
1. 12. 57nujum,
5oIbn al-Furat, IX, VI, pp. 39-40, p. 40, p. 41, 11. 18-20, pp.
p. 119, 11. 4-5; Ibn Qadi Shuhba 129-135, p. 241, 11. 13-14, pp. 246-247, p. 248, 1. 8, 11.
fol. 40A, 11. 18-19.
61 Ibn al-Furit, IX, 13-20, p. 251, 11. 7-10, p. 267, 11. 11-15. See also notes
52 Ibn
p. 129. below.
al-Furat, IX, p. 130, 11. 4-6, p. 131, 11. 16-18, 58nujfm, VI, p. 251, 11. 16-18.
p. 141, 11. 21-24. 59
53 Ibn al-Furat, IX,
Ibid., p. 260, 11. 7-8, 11. 13-14.
p. 159, 11. 11-18. 6"Ibid.,
4 Ibn al-Furat, p. 261.
IX, p. 211, 11. 4-6. 61nujfm, VI,
p. 300, 11. 15-16.
142 AYALON: The Circassians in the Mamlrlk Kingdom

other nothing is left (implying that in the natural end of the Mamluk period. Ibn Taghri Birdi's
course of events they would bring about their assertion that Barsbay, and even Tatar, were fa-
mutual annihilation). Ibn Taghri Birdi, on the vourably inclined towards the Circassians and
other hand, argues in justification of Faraj, whom secretly gave them precedence,68seems more appli-
he praises on account of his great bravery. 'For cable to an earlier state of affairs, since by the time
how is it possible, seeing that the Circassians are of these sultans Circassian supremacy was indis-
to this day the rulers of Egypt ? For Faraj killed putably established. Domination by a single race
not only them, but the members of other races as and racial solidarity are far more pronounced in
well.' 62 He also claims that Faraj only began the Circassian than in the Bahri period. Al-jins,69
slaughtering the Circassians after numerous re- meaning 'The Race,' denotes the Circassian race.
volts on their part which he had pardoned. The Similarly, al-qawm,70' The People,' is applied only
other sultans would have adopted measures of to the Circassians (see the examples given below).
similar stringency after the first revolt.63 But it is Al-ta'ifa,71 ' The Sect, Community,' also refers
evident, even from Ibn Taghri Birdi's own account, perhaps to the Circassians. On the other hand,
that Faraj's campaign of extermination was di- al-jinsiya, ' racial solidarity,' does not appear to
rected primarily against the Circassians who, al- have been in more frequent use during the Circas-
though not completely wiped out, as Maqrizl sian than during the Bahri period.72 This, how-
asserts, nevertheless received a very severe blow. ever, is probably fortuitous. Racial solidarity and
Moreover, Ibn Taghri Birdi himself relates else- hauteur characterized the Circassians while they
where that Faraj killed off countless numbers of were still a minority and challenging the suprema-
his father's Mamliuks.64 In revenge for his ex- cy of the Turk and arousing their enmity during
cesses, Faraj himself was finally killed with the the Bahri period.73 We shall discuss below the
approval of the leading Circassian amirs, with the close ties the Circassians maintained with their
exception of al-Mu'ayyad Shaykh.65 families in the Caucasus, which they brought over
The accession of al-Mu'ayyad Shaykh produced in large numbers to the Mamluik Kingdom. The
an unexpected result; for he favored the Turk importance they laid on race as the primary con-
until the majority of his amirs were of that race.66 sideration for military promotion was in evidence
According to Ibn Taghri Birdi, he resembled the already in Barqfuq'stime, as can be seen from the
former sultans in that his criterion for the choice indisputable testimony of Ibn Taghri Birdi, quoted
of soldiers was not race, but efficiency and courage.67 above. The policy of racial preference was closely
Does this imply that during the later period the maintained throughout their history. Yalkhuja
Turk were considered to be a more martial race min Mamish was unable to attain to high rank,
than the Circassians? While it is difficult to answer despite his comeliness, his great skill in games of
this question on the testimony of the contemporary furnsfya, and the fact that he was educated with
sources, there are no indications that al-Mu'ayyad 68manhal, III, fol. 11. 14-15; ibid., II, 58A, 11.
Shaykh's preference for the Turk was extended 17-8; Nujum, VI, 773, 188A, 11. 9-11.
beyond the circle of the amirs or that it led to any 69nujum, VII, p. 667, 11. 9-14. The example given by
radical changes in the army. Throughout the Popper, likawnihi turkmdniyan ghayr al-jins, means 'a
course of his reign we find no signs of racial con- Turcoman and not a Circassian' (and not as translated
flicts or tension among the Mamlfiks. by Popper in the glossary to the sixth volume of the
nujum, under jins). See also notes below, esp. note 81.
Following the convulsions during al-Nasir Faraj's 70
nujim, VI, p. 677, 1. 14, p. 688, 1. 3; ibid., VII, p.
reign and the preference evinced by al-Mu'ayyad 262, 1. 12, p. 670, 1. 10; and especially nujufm, VI, p.
Cf. 11. 10-14 and 11. 20-21.
Shaykh for the Turk, the Circassians regained 547. 71nujum, VI, p. 155. Cf. 1. 4 and 1. 13, p. 273, 1. 16.
their supremacy, which they maintained unques- See also notes below.
tioned and without any serious challenge, till the Zettersteen, Beitrdge zur Geschichte der Mamluken
72

Sultane, Leiden, 1919, p. 29, 1. 19; nujum (Cairo), VIII,

82Ibid., pp. 273-274. p. 42, 1. 8; Ibn al-Furat, VIII, p. 179, 1. 17; akhbdr
"8Ibid., p. 271, 11.4-15. timur, p. 29, 11. 13-30, 1. 2; subh, IV, p. 458, 11. 7-8;
64
manhal, IV, fol. 218A, 1. 6, 11.19-21. nujium, VII, p. 591, 1. 25; manhal, III, fol. 181B, 1. 22.
65
nujum, VI, p. 311; manhal, VIII, fol. 380B. 73
See notes above. On the hatred of the Turks for the
6 nujum, VI,
p. 430. Circassians in the Bahri period see also manhal, II,
67
manhal, III, fol. 168A, 1. 21-fol. 168B, 1. 4. fol. 96A, 11. 20-21.
 AYALON: The Circassians in the Mamlck Kingdom 143

the sons of the sultan and not in the ordinary other virtue to commend him than that he is a
military school, simply because he was a Turk and Circassian by race belonging to the Race of The
not a Circassian.74 But for this disadvantage, he People ).82 Only in isolated cases did non-Cir-
would have become atdbak al-'asdlkir(commander- cassians attain high rank, and even then it was as
in-chief).75 Aytamish, who displayed an aggressive the result of exceptional circumstances. Janibak
attitude towards the Circassians, was deserted by is shown as opposing the temporary appointment of
them, and they all went over to the sultan.76 The Jarbash al-Muhammadi as sultan on the ground
amir Khayr Bak, who was of the Abkhdzi race, that it would later be impossible to depose him in
was at variance with his Circassian Khushdtshi- favour of Janim, who was on his way to Egypt
ya.77 Of a number of amirs who attained high from Syria. This was because Jarbash was an im-
rank, it was said that their only virtue was that portant personage and min al-jins (' of The Race'),
they were Circassians.78 Particular interest at- and was brother-in-law to many of the Khushda-
taches to the following story. As he lay on his shiya belonging to the Zahiriya and Ashraflya
deathbed, al-Mu'ayyad Shaykh wanted to ensure groups. He therefore favoured Khushqadam,
the succession of his son, al-Muzaffar Ahmad. fa'innahu min ghayr al-jins ya'ni cawnahu rumi
What did he do? He appointed as atdbak Altun- al-jfins (' for he is not of The Race, that is to say,
bugha al-Qirmishi, 'for, since he was a turki al- he is of the Rami race '), since he would then be
jins, there was no danger of his seizing the throne able to get rid of him at any moment.83 Jarbash,
in place of al-Mu'ayyad's son,' 79 which was the as appears from other sources, was a Circassian by
practice of the atdbaksduring the Circassianperiod. race.84
It was indeed the case that throughout the Circas- Contemporary writers are particularly amazed
sian period no Turk ever became sultan, although by the degree of respect shown for a number of
other non-Circassian races (such as the Rumis) incompetent and frivolous Circassians, whose words
were able to gain the throne. Of Altunbugha al- were received by their fellow-Circassians-includ-
Qirmishi, Ibn Taghri Birdi goes on to relate, 'He ing the most prominent and highly-placed among
had no other fault save only that he was not of the them-as divine utterances. The best-known of
race of The People' (kdn min ghayr jins al-qawm them was al-Shaykh Sayf al-Din Lajin al-Jarkasi.
Id ghayr) i. e. he was not a Circassian.80 Taghri Although he remained a simple trooper all his life,
Birmish, the Governor of Aleppo, is said to have his fellow-Circassians wanted to elevate him to the
made no attempt to gain the sultanate li-maWrifatihi sultanate in place of Faraj. We also know of the
anna al-qawm ld yurid'inahu lidhdlika ('because series of reforms he wanted to effect in the event
he knew that The People would not have him for of his becoming sultan. The historian's comment
on this is as follows: 'That Lajin speaks such
that'). He was also aware that, if they could,
nonsense is no wonder; but what is a source of
they would not allow him to retain his governor-
wonder is that the leading members of the khdssi-
ship likawnihi turkmdnWyanghayra al-jins ('for
that he was a Turcoman not of The Race ).81 kcya and army pay attention to him.' 85 Another
man held in high esteem was Kasaw ibn ' Abdallah
Discussing Tuikh al-Jarkasi, the historian writes:
al-Zahiri, also a simple trooper and a prospective
wa-ld a'rifu fihi min al-mahdsin ghayra annahu
candidate for the sultanate.86 Apart from these
jarkasi al-jins min jins al-qawm ('I know of no
82 manhal, IV, fol. 4A, 11. 6-8.
74The education of a Mamlfik together with the s8
nujfim, VII, p. 667, 11. 9-20. Janibak used all these
sultan's sons, instead of in the school with the other arguments as a trick in order to establish Khushqadam
Mamlfiks, was considered a signal honor vouchsafed to on the throne; he succeeded only because his arguments
only a few. The small number of Mamlfiks brought up seemed to be very convincing.
in this manner succeeded very rapidly in their careers. 84
.daw', III, p. 66, 1. 11.
75 manhal, VIII, fol. 444A, 1. 10, fol. 444B, 11. 2-6. 85
manhal, V, fol. 56A, 11. 22-23; nujitm, VI, p. 155,
76
nujum, VI, pp. 18-19. 11. 3-17. Also cf. Poliak, REI, 1935, p. 244. Poliak's
77
nujum, VII, p. 857, 11. 6-7. definition of the qardnis and his description of their
78manhal, under Qirqmas al-Ashrafi; cf. nujum, VII, status in the Mamlfik Kingdom are founded on error.
pp. 24-26; manhal, IV, fol. 4A, 11. 6-8. We deal with the matter separately elsewhere. Criti-
7
nujum, VI, p. 547, 11. 10-14. cisms, largely valid, of Poliak's definition are given by
8onujtm, VI, p. 547, 11. 20-21.
Popper in his glossary to the hawddith under qirnds.
81 86 manhal, V, fol. 45A.
nujuim, VII, p. 58, 11. 14-16.
 144 AYALON: The Circassians in the Mamlakr Kingdom

there were others who gained the admiration of so effective a military training as the Mamluiks
the Circassians for no apparent reason.87 The who had been brought over at an early age. These
attitude of the Mamlfukhistorians to the members older immigrants frequently obtained high posts
of the Circassian race can be seen from a remark as amirs, or at least as kh&ssiktya,91 without being
concerning one of the Circassian amirs: wa-'indahu slaves and without undergoing training in the
taysh al-shubfubya wa-khiffat al-jarakcisa('he dis- military schools. Indeed, it would be no exaggera-
plays the giddiness of youth and the frivolity of the tion to call the second half of the Circassian period
Circassians ,).88 'the period of rule by brothers-in-law and rela-
One of the most characteristic features of the tives.' It is, for instance, related that Aynal sat
Circassian period is the practice of the sultans and securely on the throne because all the amirs were
amirs to bring over their relatives from their his brothers-in-law.92 Under Qa'itbay the rule of
country of origin in numbers unprecedented in the relatives reached its climax.93 But even in the
earlier period. Although there is no lack of evi- reign of al-Ghawri there was no lack of instances
dence that this practice was also followed in the where relatives occupied high posts.94 Maqrizi
Bahri period, it was on an incomparably smaller complains bitterly that, whereas in the past only
scale.89 The immigration of relatives reached par- young Mamluiks had been brought over, the Mam-
ticularly large proportions from the middle of the luiks in his own day had formerly been sailors,
fifteenth century onwards. The sources bearing on furnace-tenders and drawers of water in their coun-
the period in question contain numerous references tries of origin.95
to relatives being brought over, sometimes singly, The decline of the Mamlfuk Kingdom was the
but usually in groups.90 This migration had a result of a variety of causes,-political, military
number of negative features, the most important of and economic. We deal with the nature of these
which was that the immigrants included a very causes elsewhere. Even if the Circassians had not
large proportion of adults who could not be given become the dominant race during the second half
of the Mamlfik period, the cumulative effect of
87
manhal, II, fol. 129A, 11. 4-6; ibid., IV, fol. 4A, 11. these causes would finally have led to the disrup-
6-8; nujitm, VII, p. 262, 11. 12-14. tion of the Mamltuk Kingdom. It is, however,
88
manhal, IV, fol. 173B, 11. 2-3. undeniable that the Circassians appreciably has-
89
Zettersteen, p. 132, 11. 6-13; Ibn Qadi Shuhba, fol. tened this process. The later Mamliuk historians
1lOB, footnote; manhal, III, fol. 151A, 11. 14-15; durar,
III, p. 256, 11. 5-6, p. 258, 11. 10-11.
paint the early Mamlfuk period in exceedingly
90 Cf. e.g.
nujuim, V, p. 326, 11. 14-16; ibid., VI, pp. bright colors, while their account of the later
149-150, p. 519, 11. 20-29, p. 571, 11. 5-9, p. 604, 11. 8-10, period is decidedly somber. Although the account
pp. 806-807; ibid., VII, p. 93, 11. 11-12, p. 308, 1. 4, p. of the later period is to a great extent exaggerated,
465, 11. 5-6, p. 572, footnotes; hawddith, p. 38, 1. 11, p. it is basically true. Ibn Taghri Birdi's claim that
55, 11. 11-13, p. 594, 11. 13-15, p. 730, 11. 10-13; manhal,
I, fol. 9A, 11. 2-5, fol. 126B, 11. 4-7, fol. 168A, 11. 14-17, Barquq 'turned the affairs of the realm upside-
fol. 175B, 11. 12-14, fol. 203A, 11. 1-2; ibid., II, fol. 23A, down, he and his successors to this day,' contains
11. 6-9; ibid., III, fol. 177B, 11. 1-6; ibid., VIII, fol. a great deal of truth, and is only one of many
425A; tibr, p. 377, 11. 17-18; Ibn al-Furat, IX, p. 33, examples of that historian's profound insight into
11. 13-15, p. 278, 11. 23-24; Ibn Qadi Shuhba, p. 62, 11. the character of the Mamlfik regime. To translate
20-22; Ibn Iyas, II, p. 64, 11. 13-15, p. 141, 11. 18-20,
p. 210, p. 226, 11. 3-4; ibid., III, p. 78, 11. 16-18, p. 153,
11. 7-9, 22-23, p. 172, 11. 7-8, p. 173, 11. 2-7, p. 184, 11. 91nujim, VI, p. 604, 11.8-10; hawddith, p. 730, 11. 10-
17-20, p. 244, 1. 6, p. 259, 11. 4-5, p. 281, 11. 16-17, p. 291, 13; manhal, I, fol. 9A, 11. 2-5; ibid., II, fol. 175B, 11.
11. 18-19, p. 364, 11. 14-15, p. 386, 11. 21-22; ibid., IV, 12-14; ibid., VIII, fol. 425A; Ibn Qadi Shuhba, fol. 62A,
p. 88, 11. 11-13, p. 129, 11. 11, 15, p. 156, 11. 12-13, p. 184, 11. 20-22; see references in the following notes.
1. 1, p. 191, 1. 22, p. 216, 1. 17, p. 362, 1. 6, p. 433, 11. 92Ibn Iyas, II, p. 64, 11.13-15.
15-18, p. 535, 1. 10, p. 437, 11. 6-8, p. 445, 1. 20-p. 466, 93Ibn Iyas, II, p. 210, p. 226, 11.3-4; ibid., III, p. 78,
1. 2, p. 477, 1. 21; al-Ish.aqi, akhbdr al-uwal, Cairo, 1315 11. 16-18, p. 153, 11. 7-9, 22-23, p. 172, 11. 7-8, p. 173, 11.
A. H., p. 124, 11. 21-24; daw', II, p. 315, 1. 23; ibid., III, 2-6, p. 184, 11. 17-20, p. 244, 1. 6, p. 259, 11. 4-5, p. 281,
p. 28, p. 63, 1. 8, p. 284, 11. 17-18, p. 287, 1. 23; ibid., VI, 11. 16-17, p. 291, 11. 18-19, p. 364, 1. 14-15, and in many
p. 201, 1. 1, p. 219, 11. 2-4, p. 221, 11. 1-3, p. 271, 1. 1; other places.
ibid., X, p. 288, 1. 24; zubda, p. 121, 11. 2-4; see also 94 Ibn Iyas, IV, p. 129, 1. 11, p. 156, 11. 12-13, p. 184,
Poliak, REI, p. 242, Note 4, also material collected by 1. 1, p. 191, 1. 22, p. 216, 1. 17, et passim.
Poliak on this question, REI, 1935, p. 242, Notes 3, 4, 5. 65khitat, II, pp. 213-214.
 AYALON: The Circassians in the Maml,uk Kingdom 145

Ibn Taghri Birdi's judgment into more scientific pass through the crucible of the military school,
language, one might say that the Circassians struck where they were turned from infidels into Mus-
a severe blow at several of the main supports on lims, from boys into grown men, from raw recruits
which the military system of servitude in Islam into full-fledged troopers, and from slaves into
was founded. From the establishment of the first free men, the result would have been to deprive the
Mamliuk companies in the ninth century in the Mamliukarmy of its very lifeblood. (Nevertheless,
Abbasid Caliphate down to the decline of the the majority of the Mamliuksin the late Circassian
system of servitude under Ottoman rule, several period did pass through the school, but their
fundamental principles were in operation: training was extremely perfunctory in comparison
with earlier times.) Presumably there was not a
(a) A protracted period of training for the total absence of nepotism, rapid promotions and
Mamluiks in the military school from boyhood. racial favoritism in the earlier period; but these
This training formed the character of the Mam- practices certainly had not assumed the vast
pro-
luiks and determined their outlook for the rest of portion they did in Circassian times.
their life. One of the main reasons for the amazing sta-
(b) Blind obedience by the Mamliukto his over- bility and compactness of the Ottoman slave family
lord, and unwavering loyalty to his companions in (down to the end of Suleiman the Magnificent's
servitude. reign) 97 was the fact that in the Ottoman State
(c) The importance attached to ability as the the sultanate was hereditary within the reigning
chief criterion of promotion. family, and often passed from father to son. Thus
the allegiance of the slaves was transferred from
(d) A very slow rate of promotion. one sultan to his successor. With the Mamluks,
(e) The unusual degree of importance attached however, the
legitimacy of kingship had become
to senior Mamluiks as contrasted with younger weakened
already in Bahri times. Even during
ones. the rule of the house of Qaltaun, which lasted
Even at the most flourishing period of the Mam- longer than any other Mamluk dynasty, Muham-
luk Kingdom, these principles were not elaborated mad b. Qalaun was on two occasions forced to
to the degree of perfection witnessed under Otto- vacate the throne to make way for powerful Mam-
man rule. Nevertheless, they remained the back- luk amirs. During the Circassian period the prin-
bone of the system of military servitude in the ciple of heredity was altogether abandoned (see
Circassian period as well. But at the same time above), so that it was impossible for the Mamluks
there is no doubt that they received considerable of one sultan to transfer their allegiance to his
modification at the hands of the Circassians in successor. At this time in particular, the Mamluks
contrast to the earlier period. It is not our inten- purchased and liberated by the reigning sultan (al-
tion here to furnish a full account of the manner mushtarawat, al-ajlab, al-julbdn; see note 22)
in which the foundations of the Mamluiksystem of formed the entire basis of his rule, and were the
servitude were weakened during the Circassian sole support on which he leaned in opposing the
period. We should, however, like to review in brief other groups of royal Mamlufks,98and especially
certain aspects of that process. The military school the Mamliuksof his predecessor. The sultan and
was adversely affected in two ways: the period of his mushtarawdt formed a sort of vicious circle.
study was considerably curtailed as compared with They enjoyed pre-eminence only as long as he
the early Mamliukperiod; 96 and, as we have seen, reigned; and he could retain his power only as
a large proportion of adults-relatives of the Cir- long as he based his rule on his Mamluks. Each
cassian amirs-were admitted to the Mamlufikcorps 97The causes of its decline after
the reign of Suleiman
without passing through the school at all. This have been dealt with in detail
by A. H. Lybyer, The
practice was completely at variance with the prin- Government of the Ottoman Empire in the Time of
ciples underlying the system of military servitude. Suleiman the Magnificent, Harvard University Press,
If it had become the universal practice for Mam- 1913. Lybyer gives in this book a most illuminating
of the foundations on which the military
luks to arrive in Egypt as adults and no longer to description
slave system was based, both in the Ottoman Empire and
96
in Islam in general.
nujum, VI, p. 509, 11. 15-19. 98 For the structure of the Mamluik army see Note 22.
146 AYALON: The Circassians in the Mamllk Kingdom

sultan, on ascending the throne, tried to clear the tions of the formation and subsequent disruption
way for his mushtarawdtto seize power, and sought of temporary alliances among the qardnis. These
to increase their number as rapidly as possible. alliances were mostly of a very casual nature, and
The accession of a new sultan was frequently ac- the stability of the sultan's rule depended largely
companied by brutal purges on a vast scale, and upon his ability to exploit the differences among
especially by the relentless persecution of his pre- the various groups of Mamlfiks in his own interest.
decessor's Mamluiks. In other words, he would try As long as the Mamliuk Kingdom was wealthy
to eliminate the Mamlfukswho had only just before and powerful, and its army well-trained, disciplined
been the mushtarawdt, and thus had been the rul- and controlled by an iron hand, the feeling of
ing element in the state. In order to destroy the comradeship between the slave and liberated Mam-
immense power they had concentrated in their luks of the same overlord (khushdashiya) was a
hands, and to weaken their opposition as a united positive factor. There was a restrained competi-
and organized body, the sultan was not content tive feeling, an esprit de corps, among the various
merely to remove them from their dominant posi- military formations, which was useful in keeping
tion. In addition he would employ far more the army up to the mark and preventing it from
stringent measures, throwing some of them (mainly growing stale. When, however, the MamliukKing-
the leaders) into gaol or exiling them to remote dom began to decline under the Circassians, and
parts of the kingdom, and transferring others to discipline was relaxed and the egoistic passions of
the service of the amirs, which was considered as a the different formations were permitted to break
great degradation. out unchecked, the spirit of comradeship proved
This practice was a feature of Mamlfik rule harmful to the army and was one of the causes
from its earliest days; but in the Bahri period it of its disintegration. The way was opened to com-
was carried out in a far milder fashion than under plete domination by the mushtarawdt of the other
the Circassians. For in the Bahri period the prin- military groups, accompanied by unlimited extor-
ciple of hereditary succession had not altogether tion and oppression. It was also made possible for
been abandoned, and the sultan's son could not the soldiers to occupy themselves constantly with
regard his father's Mamliuksas complete strangers. political affairs, to the neglect of their military
Similarly, the father's Mamlfiks could not regard duties. Matters reached such a pass that the sultan
his son as a completely strange sultan. would send the qaranis-the members of the
Opposed to the mushtarawdt were the Mamluiks veteran formations-on military expeditions, so as
of the preceding sultans (al-qardnisa, al-qardn.s). to spare his own mushtarawdt,99whose military
Unlike the mushtarawat, who were a homogeneous training was of a very low order and who were not
body, the qardnis, of their very nature, could not moved by a martial spirit. According to Ibn
be a compact group; for they were composed of Taghri Birdi, a hundred Mamlfiks of the qardnis
heterogeneous, and frequently mutually antagonis- were capable of putting to flight more than a
tic, elements owing allegiance to different ex- thousand julbdn.100But if they did not show much
sultans and drawn together solely by their com- prowess in the field, they exhibited an amazing
mon hatred of the younger mushtarawdt, who aptitude for political intrigue, overcoming and
sought to dominate them. Thus the Mamlfiks of suppressing the weaker military formations, and
ex-sultan A could not forget that the Mamluiksof indulging in acts of wanton lawlessness. Signs of
ex-sultan B had ousted them from their former po- the weakening of discipline had been manifested at
sition as mushtarawdt; and the Mamluks of ex- the beginning of the Circassian period, and there
sultan C were hostile to those of ex-sultan B for the had even been sporadic manifestations during
same reason; and so on. The possibilities of fomen- Bahri times; but from the beginning of the second
ting discord among the qardnis were particularly half of the ninth century A. H. all barriers were
extensive during the Circassian period, partly for broken down. The historians point to Aynal's
the reason that the short reigns of the Circassian 99 This subject is dealt with at length in the chapter
sultans resulted in the formation of many such entitled, ' The Army on the March,' in our work on the
groups. Thus plentiful opportunities were pro- Mamlfuk army.
vided for creating various combinations and align- 00??nujm, VI, p. 641, 11. 2-5. See also the corrigenda
ments. The Mamlfik histories are full of descrip- on p. LXXII of the volume in question.
 AYALON: The Circassians in the Mamlik Kingdomn 147

reign as the time when all restrictions on the wild largely favored by the fact that, from the time of
behavior of the julbdn were swept aside. Describ- Timuir Lang, on the threshold of the 15th century,
ing the year 858 A. H. Ibn Taghri Birdi writes: down to the war of Dabiq at the beginning of the
' This is the first appearance of the al-Ashraf 16th, the Mamliuks did not engage in a single
(Aynal) Mamlfiks; and what follows is still more large-scale war. A serious menace of external ag-
terrible.' 101 From the middle Qf the ninth century gression, such as the Mongol threat during the
A. H. the Mamluk chronicles bear the impress of second half of the 13th century, would either have
the reign of terror of the julbdn, who completely led to the cohesion of the Mamluk slave family or
dominated all the other groups of slaves and, in to the disruption of the kingdom (as actually hap-
fact, all aspects of life in the Mamlfik Kingdom. pened in 922 A. H.). The absence of any such
This subject, indeed, engrosses the attention of the threat during the greatest part of the Circassian
chroniclers during the last 70 years of the Mamlfik period promoted the process of disintegration and
regime.?02 increased the differences among the Mamlufks.
The incessant conflict between the various groups Continued pressure on the Ottoman slave regime,
of royal Mamlfuksduring the Circassian period was which was constantly engaged in fighting difficult,
101nujum, VII, p. 452, 11. 6-7. protracted, and bloody wars-firearms were largely
102
The conclusions given here in brief concerning the
employed in them-was without doubt one of the
adverse effect of the Circassians upon the foundations reasons for its effectiveness, cohesion, and suprema-
of the Mamlfik system of servitude, the relations between cy. But the similarities and dissimilarities between
the various Mamlufik formations, etc., are based on the Ottoman and Mamluk slave systems, which
material compiled in the chapter entitled 'The Structure
of the Mamlfik Army' in our work on the Mamlfik army.
form one of the most important aspects of the
Notes 96 and 98-101 give only some of the more history of Islam in the late Middle Ages, calls for
important references. separate treatment.