The Berbers of the Arabs

Author(s): Ramzi Rouighi

Source: Studia Islamica , 2011, Vol. 106, No. 1 (2011), pp. 49-76
Published by: Brill

Stable URL: http://www.jstor.com/stable/23884946

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Brill is collaborating with JSTOR to digitize, preserve and extend access to Studia Islamica

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
BRILL Studia Islámica 106 (2011 ) 49-76 brill.com/si

The Berbers of the Arabs

Ramzi Rouighi
University of Southern California

Moors and Berbers

The Arab conquest of northwest Africa is a major event in the field ofNorth
African Studies. For scholars, its most lasting effects were the Islamization
and Arabization of the Berber populations. While historians often disagree
about the chronology, character, and extent of these processes, they ail
agree that they are essential to understanding the medieval period.
In truth, the scholarly consensus is slightly more complicated than this
more "popular" view. Specialists know that the category "Barbar" is of Ara
bie origin and thus that the Arabs could not have conquered peoples called
Barbar prior to the conquest.1 Furthermore, while Latin and Greek sources
refer to some people in northwest Africa as barbarians (βάρβαροι and bar
bari), the ideas associated with these barbarians wil be shown to be so dif
férent from the Arabie Barbar that it is difficult to confuse the two. Thus, in
spite of the obvious linguistic similarity of the Greek, Latin, and Arabie
terms, the Arabs did not apply the term Barbar to describe the exact same
groups of peoples called "barbarians" in classical and late antique sources.
So, who were the Barbar of the Arabs?
Encapsulating a widely held view among specialists, Gabriel Camps
explained, "The Berbers of the Arabs are the Moors of the Romans."2
According to this understanding, instead of conquering the Berbers, the

1 "Berber" is the most common Romanization of the Arabie word. The more
scholarly translitération is "Barbar." The term Barbar in this essay highlights the
modernity of "Berber," the category most scholars have used to produce pre
Islamic Berbers.

2 Gabriel Camps noted that "En fait les Berberes des Arabes sont les Maures des
Romains," in Paul-Albert Février, La Méditerranée de Paul-Albert Février, 2 vols.,
(Rome: Ecole Française de Rome, 1996), 2:305. The idea has permeated scholarship.
Koninklijke Brill NV, Leiden, 2011 DOI: 10.1163/19585705-12341252

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
50 R. Rouighi / Studia Islámica 106 (2011 ) 49-76

Arabs conquered the "Moors" (Mauri, sing. Maurus) and just called them
by a "new" name. The historian's task was simple: analyze the meaning of
"Moor" in those Greek and Latin sources written just before the Arab con
quest and compare them to the meanings of "Barbar" in the earliest Arabie
sources. Unfortunately, Latin and Greek sources on North Africa begin to
diminish in the sixth Century, while Arabie sources on the "Barbar" only
appear in the ninth Century. In some respects, the équation "Moor = Bar
bar" attempts to résolve the problem posed by the record.3
Whether the sources call them Moors or Barbar, some scholars believe
the people were the same.4 Proceeding from such a presupposition, histo
rians have grown accustomed to using the category "Berber" to refer to
northwest Africans in general, in any period. For these scholars, the Berbers
were the same indigenous people of northwest Africa conquered by the
Arabs, Byzantines, Vandals, Romans, and Phoenicians.5
In his recent study of the "Moors," historian Yves Moderan thoroughly
examined the use of the term in late antique sources.6 His goal was to deter
mine the role of the populations named Moors then Berbers in the évolu
tion of Roman Africa in the three centuries preceding the Arab conquest.7
Noting the difficulties inherent in the use of "Berber" to analyze the pre
Islamic period, Moderan commented on the scholarly practice of translat
ing "Moor" as "Berber":

For a recent illustration, see Eduardo Manzano Moreno, Conquistadores, Emires Y

Califas:Los Omeyasy Laformación de al-Andalus, (Barcelona, 2006), 29.
3 For a thorough discussion of late antique sources, see Yves Moderan, Les
Maures et l'Afrique Romaine {TVe-Vlle siècle), (Rome, Ecole Française de Rome,
2003).
4 Historians sometimes extend the équation and use the term "Berber" to refer
to various ancient peoples in the région. For a recent example, see Elizabeth Fen
tress, "Romanizing the Berbers," Past and Présent, 190 (Feb. 2006), 3-33.
5 See for example, A.H. Merills (ed.), Vandals, Romans, and Berbers: New Per
spectives on Late Antique North Africa, (Aldershot, 2004); David L. Stone and Lea M.
Stirling, "Funerary Monuments and Mortuary Pratices in the Landscapes of North
Africa," in idem (eds.), Mortuary Landscapes of North Africa, (Toronto, 2006), 3-31.
6 Since Arabie sources had little use for the category "Moor," Moderan's study
effectively ends in the seventh Century. His analysis of Arabie texts is problematic
because of its liberal use of much later medieval sources such as Ibn Khaldün
(d. 1406). Additionally, Moderan utilizes dated French translations, which intro
duce decidedly modem concepts such as "race."
7 Yves Moderan, Les Maures, 811. Moderan's study is both a recent and seri
ous articulation of this position. In addition to critical historiographie essays, it
includes a rich bibliography, which lists ail the major works on this issue.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011)43-76 51

[It] imposed itself immediately to the first historians of late antique

Africa without any justification, and has remained unchanged until the
présent. In fact, as we shall see, if in the sixth Century the Moor was still
an indigenous (or autochthonous) African (autochtone) whom the
Romans considered to be "non-Romanized," he could possess a cultural
complexity that is far from the "pure Berber" dear to nineteenth-cen
tury scholars.8 In order to respect the sources, and avoid any anachro
nism and all ideological ambiguities [inherent] in the word "Berber,"
we will mostly speak here of "Moors." But as we will see, obvious stylis
tic reasons make it so that we could not avoid "Berber" in some sen
tences.9

In spite of Moderan's anachronistic and ideologically ambiguous use of

Berber "in some instances," it is important to note that his ability to do so
rests on the idea that both Moors and Berbers were "indigenous."10
In this essay, I will reassess the validity of the équation of Berbers with
Moors, which underlies scholarly research, by focusing on the représenta
tion of the Barbar in early medieval Arabie sources. In particular, I will con
céntrate on the extent to which these sources conceived of the Barbar as
the indigenous inhabitants of the région. In this way, I hope to outline
research possibilities about the Arab conquests, conversion to Islam, and
Arabization.11

8 The idea of a "culturally complex" Moor does nothing to elimínate the idea
of the "pure Berber."
9 Moderan, Les Maures, 11.
10 For the Berbers as indigenous North Africans, see for example, Jamil M.
Abun-Nasr, A History of the Maghrib, (Cambridge, 1975), 1; Michael Brett and
Elizabeth Fentress, The Berbers, (Oxford & Cambridge MA: Blackwell, 1996), 10;
Maribel Fierro, 'Abd al-Rahman III: The First Cordobán Caliph, (Oxford, 2005), 11;
Helena de Felipe, Identidad y onomástica de los bereberes de al-Andalus, (Madrid,
i997)i 18; Heinz Halm, Das Reich des Mahdi: der Aufstieg der Fatimiden (875-975),
(Munich, 1991), 95-6; Eduardo Manzano Moreno, Conquistadores, Emires Y Califas:
Los Omeyasy laformación de al-Andalus, (Barcelona, 2006), 29.
11 The scholarship on these subjects is dense and I can only suggest a few nota
ble interventions. See Richard Bulliet, Conversion to Islam in the Medieval Period:
an Essay in Quantitative History, (Cambridge MA: Harvard University Press, 1979);
Patricia Crone, Slaves on Horses: the Evolution of the Islamic Polity, (New York: Cam
bridge University Press, 1980); Fred Donner, Earty Islamic Conquests, (Princeton:
Princeton University Press, 1981); Thomas F. Glick, Irrigation and Society in Medi
eval Valencia, (Cambridge MA: Belknap Press of Harvard University Press, 1970);

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
52 R. Rouighi / Studio Islámico 106 (2011)49-76

Pre-Islamic Barbar?

Did pre-Islamic Arabie authors employ the category "Barbar"? If they did,
did it correspond to the Roman "Moors"?
The first Century author of the Greek PeripLus of the Erythraean Sea
described a région named "Barbaria" south of the Egyptian town of Bere
nike on the Red Sea coast.12 A Century later, the famous geographer Clau
dius Ptolemaeus (ca. 90-ca. 168) located the Barbarían Sea across from the
Bay of Arabia, the Red Sea, and the Barbarían Bay. The latter led to the city
of Rhapta, which he described as the metrópolis of Barbaria.13 In the
Períplus of the Outer Sea, Marcianus of Heraclea Pontica (fl. 400) used the
ñame to describe an ethnos and a sea.14 In the sixth Century, Cosmas Indi
copleustes, famed for his sea travels to India, located Barbaria right across
the Arabian Gulf, as did his contemporary, Stephanus of Byzantium.15
Clearly, for six centuries Greeks and Romans consistently and regularly
described a Barbaria on the east coast of Africa.
The pre-Islamic Arabs also knew about this Barbaria and Barbar on the
other side of the Red Sea facing Arabia. Excerpts from early authors, known
through later recensions, contain references to them. The famous poet
Imru'u al-Qays (6th c.) mentions the Barbar and their horses in one of his

Walter Kaegi, Byzantium and the Early IsLamic Conquests, (New York: Cambridge
University Press, 1992); Michael Morony, Iraq after the Muslim Conquest, (Prince
ton: Princeton University Press, 1984); Chase Robinson, Empire and Elites after the
Muslim Conquest: the Transformation of Northern Mesopotamia, (New York: Cam
bridge University Press, 2000).
12 Periplus Maris Erythrcei, in Geographi Graeci Minores, ed. Karl Muller, (Paris:
Firmin Didot 1855), 1:258.
13 Claudii Ptolemaei, Geographia, ed. C.F.A. Nobbe, (Hildesheim: G. Olms,
1966), 1,17,6; IV, 7,28. This Barbaria was thus located on the Red Sea coast of mod
ern Somalia.
14 Marciani Heracleensis ex Ponto, Periplus Maris Exteri, in Geographi Graeci
Minores, ed. Karl Muller, 1: 523.
15 Cosmas Indicopleustes, Topographie Chrétienne, ed. Wanda Wolska-Conus,
(Paris: Editions du Cerf, 1968), 11:26; 11:2g; 11:30; 11:45! 11:48; 11:49! IL50; II:6i; 11:64.
Stephani Byzantii, Ethnicorum Quae Supersunt, ed. A. Meineke, (Berlin, 1849), 158.
See "Barbaria" in Georg Wissowa, Paulys Real-Encyclopädie der classischen Alter
tumswissenschaft, (Stuttgart: J.B. Metzler, 1896). In addition to the east African
Barbar, one notes the existence of a Barbar in the Persian Gulf. See the suggestive
article by G.W. Bowersock, "Tylos and Tyre: Bahrain in the Graeco-Roman World,"
in Bahrain through the Ages: the Archaeology, Shaikha Haya Ali Khalifa and Michael
Rice eds., (London: Routledge & Kegan Paul, 1986), 399-406.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 49-76 53

poems.16 Another pre-Islamic poem attributed to 'Uday b. Zayd (d. 587)

mentions in the same line the people of Barbar {al barbar) and al-Yaksùm
(Axum).17 This use of Barbar corresponds to the one found in Graeco
Roman géographie texts. It does not refer to a particular people like the
Arabs or the Abyssinians but rather to the inhabitants of a région or city on
the east coast of Africa called Barbar.18
Furthermore, the historian Ibn 'Abd al-Hakam (d. 871) mentioned a Bar
bar market {süq Barbar) in the Egyptian city of Fustät.19 The absence of the
definite article indicates that it was the market of the people of Barbar {äl
barbar) not the Barbar {al-barbar). Again, the Barbar in question is the east
African région to the south of the city of Fustät. Thus, one may safely con
clude that the name predated the Arab conquest of northwest Africa.
Unlike the barbarians of ancient and late antique sources, which refer to
"uneivilized" peoples found just about anywhere in the known world, the
Barbar of the pre-Islamic Arabs appear in a specific geographical area. Thus,
it is safe to conclude that in pre-Islamic times, the Arabs referred to some
people as "Barbar" and that they imagined them as the inhabitants of east
ern, not northwestern, Africa. Since these Barbar were not also Moors, it is
necessary to find out whether the Arabs gave up this older conception and
took on a new one once they invaded northwest Africa.

Conquest and Politics

The earliest Arabie sources describing the Muslim conquest of northwest

Africa and al-Andalus were composed in the ninth Century, more than a
Century after the first raids. In these texts, "Barbar" refers to northwest

16 Medieval authors made repeated references to this fact. For example, see Ibn
Khaldün (d. 1406), The Muqaddimah, an Introduction to History, Franz Rosenthal
trans., 3 vols., (New York: Pantheon Books, 1967), 1:99.
17 Ibn Hisham, Das Leben Muhammed's nach Muhammed Ibn Ishäk, (Kitäb slrat
rasül Allah), ed. Ferdinand Wustenfeld, (Frankfurt am Main: Minderva, 1961), 45.
Note that in his translation of the work A. Guillaume incorrectly translates "Ä1
Barbar" as barbarians. See A. Guillaume, The Life of Muhammad, 2nd ed., (Lahore,
1967). 33
18 See al-BukhärT, Ta'rlkh al-rusutwa al-mulük, MuhammadAbü al-Fadl Ibrähim
ed., 11 vols., (Cairo: Dar al-Ma'ärif, i960), 8: 450. In an enumeration of peoples, the
Barbar are situated between the Nüba and al-Südän.
19 Ibn 'Abd al-Hakam, The History of the Conquest of Egypt, North Africa, and
Spain, Charles Torrey ed., (New Häven: Yale Oriental Series, 1922), 111,115,119.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 54 R- Rouighi / Studia Islámica 106 (2011) 43-76

Africans. As they tend to take the category for granted, these texts need to
be critically reassessed in light of the political events they recount. A synop
tic narrative of the Arab conquest of northwest Africa and the politics that
ensued will set the stage for a discussion of early Arabie sources and their
portrayal of the Barbar, and how the term, originally applied to east Afri
cans, came to be applied to northwest Africans.
The Arab conquest occurred in stages.20 In the 640s, Arab generáis based
in Egypt led armies westward in search of booty and honors, and initiated
a graduai conquest that lasted decades. Over time, the incursions, raiding,
and Settlements led to a slow and progressive reorientation of politics in
northwest Africa, closely related to the new important centers of power in
the Mashriq. Political struggles taking place in the east had a great impact
on the formulation of a course of action, the leadership of the military cam
paigns, as well as shaping the terms of the opposition. The garrison town of
al-Qayrawân, established in 670, became the capital of Arab presence in
the west.

The articulation, and thus terminology, of political struggle between

Muslims in the Maghrib owed a great deal to developments in the Mashriq.
In this regard, it may seem as if eastern politics were transplanted in the
west, or that the west reproduced the politics of the east. Yet, and this is
crucial, the parties that won battles in the east did not necessarily or auto
matically win in the west as well.
By the early eighth Century, the Arabs had ended eastern Roman rule in
Ifrlqiyä and more or less controlled that région. The masters of al-Qayrawân
tried to expand their control beyond Ifrlqiyä. As they raided, imposed trib
utes and taxes, and accumulated captives, the Arabs came up against a
number of political formations, which had emerged from the reordering of
power subséquent to the defeat of the Byzantines. The ability of these
groups to raise armies made them formidable opponents and useful allies.
After a protracted series of battles, intrigues, alliances, victories, and rever

20 My description of the events owes a great deal to the work of Jamil Abun
Nasr, A History ofthe Maghrib in the Islamic Period, (Cambridge & New York: Cam
bridge University Press, 1987) and A History of the Maghrib, (Cambridge & New
York: Cambridge University Press, 1975); Michael Brett and Elizabeth Fentress,
The Berbers, (Oxford & Cambridge MA: Blackwell, 1996); Hugh Kennedy, Muslim
Spain and Portugal: A Political History of al-Andalus, (New York: Longman, 1996);
and Abdallah Laroui, L'Histoire du Maghreb: un essai de synthèse, (Paris: Maspero,
1970).

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouíghi / Studia Islámica 106 (2011 ) 49-76 55

sais of fortune, some northwest Africans, mostly from areas near Ifrlqiyâ,
jolned with the Arabs of al-Qayrawän. Although their social background
is not known, many of them became clients or mawàlî (sing, mawlä) of
Arab leaders, a status that gave them new political and legal standing.
These clients fought alongside the Arabs and came to espouse similar
eastern idéologies.
In the first decade of the eighth Century, an Arab force reached the shores
of the Atlantic. Arab forays into the pre-Saharan and Saharan régions were
limited and, in any case, not as spectacular as their capture of the Mediter
ranean city of Tangier.
In Damascus, the Umayyads (661-750) chose this juncture to make the
Maghrib into a single province (wiläya). For the first time, the entirety of
northwest Africa became an administrative unit. This act of government
made "the Maghrib" more than just a general term for the géographie
west.

In 711, the new governor of Tangier, the mawlä Täriq b. Ziyäd, crossed
the Mediterranean into Iberia with an army predominantly composed of
northwest Africans. A few months later the Arab Müsä b. Nusayr followed
him and took over the command of the Muslims there. In subséquent
years, they accumulated further victories and pushed their raids far into
the northern régions. Al-Andalus was born.
In the west, the Arab elite put in place a System of precedence that guar
anteed them preferential treatment. The Arab elite fought hard to establish
and maintain the mechanisms that distinguished between Muslims, and
which alienated their relatively recent northwest African allies.21
By the 730s, many groups challenged Umayyad rule in the east and the
west. The grievances of opponents included the nepotism, rapaciousness,
and arbitrary brutality of Umayyad officiais. In this regard, the discontent
of those who had reasons to expect something from the government in
al-Qayrawän did not differ greatly from that of non-Arab, and Arab,
Muslims in other régions. In 739 and 740, rebellions from al-Qayrawän to
Tangier seriously threatened Umayyad rule in the région. Unsurprisingly,

21 The sources' distinction between northern and southern Arabs ('Adnânï v.

Qahtânï) is complicated by the pre-Islamic migration and settlement of Yemenis
in Syria (al-Shâm), and the later migration of southerners following the Muslim
conquest.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
56 R. Rouighi /Studia Islámica 106 (2011 ) 49-76

anti-Umayyad Arabs, some of whom had fled Umayyad police in the east,
formed alliances with various groups of northwest Africans.
In the early 740s, anti-Umayyad movements seriously confronted the
Arabs of al-Andalus. However, unlike the Maghrib, al-Andalus was a more
recently conquered territory. Because the leaders of the conquest were
Umayyad generáis, Arab elites of al-Andalus were predominantly pro
Umayyad. This did not prevent the articulation of anti-Umayyad sentiment
in ways that echoed eastern politics. It did mean, however, that the most
serious threat to their prédominance came from an alliance of anti
Umayyad Arabs and northwest Africans. For the most part, and given the
overwhelming numerical superiority of the Africans among the discon
tented, the latter led the uprisings in al-Andalus.
In al-Andalus, a similar rébellion in 741-2 showed that disunity among
Arabs could seriously challenge the status quo. Inter-Arab strife was
resolved by the landing of an army of Syrians (Shämiyyün). An Umayyad
inflected Arabism overwhelmed the rivalries and jealousies between north
ern and southern Arabs, newcomers and "older" families. In this respect,
al-Andalus was différent from the rest of the Maghrib where Islam, rather
than Arab prédominance, was the more dominant ideology.
When the 'Abbâsids and their supporters in the east put an end to
Umayyad rule in 750, al-Qayrawân hailed the new rulers. Al-Andalus took
another route. The pro-Umayyad camp welcomed the arrivai of the sole
surviving member of the Umayyad dynasty and in 756, after routing a vigor
ous opposition, declared an independent Umayyad emirate in al-Andalus.

The First Western Barbar?

The Arab-Byzantine Chronicle ofj4i22 is the earliest extant historical source

written after the Muslim conquest of the Maghrib and al-Andalus. Its
author was Iberian, and its language was Latin. Unfortunately for historians
of the Maghrib, the anonymous author was more interested in Byzantine
imperial politics and barely mentioned events in al-Andalus. One of his
contemporaries living in al-Andalus possibly used the same sources to pen

22 Juan Gil, Corpus scriptores muzarabicorum, 2 vols., (Madrid: Instituto Anto

nio de Nebrija, 1973).

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 49-76 57

a richer chronicle.23 This author of The Chronicle 0/754, also anonymous, is

more informative and describes political events in Iberia from 611 to 754.24
Although neither of these authors refers to the Barbar, the author of the
Arab-Byzantine Chronicle refers to the Moors (Mauri) once to describe the
battle that ended Byzantine rule in Africa (698).

Therefore the confrontation was prepared, whereupon, the battle line

of the Moors turned in flight and ail the nobility of Africa, along with
count Gregory, was destroyed to the point of extinction 25

The description of the Moors is rather vague and does little to identify
them. The author described the soldiers as "Moors," and distinguished them
from the "nobility of Africa."26
In contrast to the imprecisely defined Moors, the authors refer to the
inhabitants of the old Roman province of Africa as Africans (.Africani).27
Arabie authors also noted their distinctiveness and referred to them as
Afäriq. As they integrated into the Muslim polity, they came to constitute a
distinctive social group in the new Ifrlqiyä. However, the continuity in the
usage of this category from Greek and Latin to Arabie does not necessarily
mean that the Afäriq were the same families that had been considered Afri
cans (Africanii) under Roman rule. As the Chronicle indicates, the "old"
Africans were "destroyed to the point of extinction." Consequently, the
Africani of the mid-eighth Century translated the Arabie "Afäriq," which
referred to a sociopolitical group specific to Arab domination in Ifrlqiyä.

23 See Robert Hoyland, Seeing Islam As Others Saw It, (Princeton NJ: Darwin
Press, 1997), 611-630. In addition to translating the chronicle, Hoyland Highlights
parts that correspond to earlier texts he identifies.
24 Crónica Mozárabe de 754, ed. Jose Eduardo Lopez Pereira, (Zaragoza: Anubar
Ediciones, 1980). An English translation is available in Kenneth B. Wolf, Conquer
ors and Chroniclers of Earty Medieval Spain, (Liverpool: Liverpool University Press,
1990), 111-58.
25 §24: Hoyland, Seeing Islam, 618; Juan Gil, Corpus, 1:10.
26 It fits the general use of the term in the late antique period. See Yves Mod
eran, Les Maures, 11,448-50.
27 27. In the late antique period, "Africa" was applied to the province of Africa
Proconsularis, or to the area encompassing Proconsularis, Byzacena, and Tripoli
tania prior to Diocletian's reforms in the third Century. See Moderan, Les Maures,
169.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
58 R. Rouíghi / Studia Islámica 106 (2011 ) 49-76

While it is not clear if the author of the Arab-Byzantine Chronicle col

lected any information in Iberia, his contemporary, the author of The
Chronicle of754, did. In fact, he seems to have used even Arabie sources.28
This makes his chronicle the more important as it relies on Arabie sources
now lost.

A passage about the rébellion against Arab rule in both the Maghrib and
al-Andalus in 741-2 suggests that the author may have gleaned more than
dates and names from Arabie sources. For instance, he refers to "the west
ern région," which is the translation of the Arabie "Maghrib." In addition, he
describes the Moors as "naked, girded only with loin-cloths covering their
shameful parts." When they battled the Arabs, "the Egyptian horses imme
diately recoiled in flight, as the Moors on their beautiful horses revealed
their répulsive color and gnashed their white teeth."29
This story offers a very dramatic préludé to the landing of the Syrians in
al-Andalus and explains their zeal and expertise in crushing the rébellion
there. The description of the events in al-Andalus possesses concreteness
and specificity missing from the narration of African affairs. This suggests
that the Arabie sources in question were either Andalusi or had an Andalusi
connection. For instance, as he mentions Balj's "good lineage," the author
not only demonstrates intimate knowledge of politics in al-Andalus at the
time of the Umayyad victory, but also positions himself in relation to it.
Distinguishing himself from the author of the Arab-Byzantine Chronicle,
this author provides much more information about politics in Iberia. More
importantly for the purposes of this essay, he introduces two new catégo
ries: the "Moors of Spain" and the "Saracens of Spain."
The Moors and Saracens of Spain refer to groups who had privilèges as
Muslims that set them apart from Christians:

A Saracen by the name of Yahya succeeded at once by orders of the

princes. He was a cruel and terrible despot who raged for almost three
years. With bitter deceit, he stirred up the Saracens and Moors of Spain

28 Wolf, Conquerors, 29-30.

29 §84: Wolf, Conquerors, 147-8; Lopez Pereira, 107-8. Lopez Pereira's translation
introduces the concept of the race of the Arabs (la raza de los árabes) which does
not appear in the Latin text. For Ibn 'Abd al-Hakam (d. 871), the episode is told in
terms of the struggles between Arab generáis which resulted in the poor décision
to send Arab horsemen against naked Barbar foot-soldiers. The "Egyptian" horses
are the horses of Kulthüm the general who arrived from Egypt. The naked Barbar
are described as "Sufrí" khärijites. See Ibn 'Abd al-Hakam, 219-20.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 4g-/6 59

by confiscating property that they were holding for the sake of peace
and restoring many things to the Christians.30

However, the Moors of Spain became threatening when their ties to "their
people" in Africa led them to revolt against the Saracens of Spain.

Although he was prééminent in courage and famé, a Moor named

Munnuza, hearing that his people were being oppressed by the harsh
temerity of the judges in the territory of Libya, quickly made peace
with the Franks and organized a revolt against the Saracens of Spain.31

The anonymous author cast the revolts of the 730s and early 740s in terms
that mark them as a traumatic experience. Inter-Arab strife had prevented
Umayyad troops from Coming to the rescue and encouraged the Moors of
Spain to rebel endangering the status quo.

When [in 742, 'Abd al-Malik] discovered that the third part of the army
under Balj had arrived at the port, he denied them a crossing, withhold
ing the ships. When the Moors of Spain realized this, they assembled
for war, wanting to subject 'Abd al-Malik to themselves and, crossing
over the sea in ships, offer his conquered kingdom to their allies on the
other side of the sea.32

Clearly, the rebellions of the Moors against the Arabs in al-Andalus pro
duced conditions requiring a distinction between the Moors of Iberia and
those of northwest Africa. In this regard, one may see the generic use of the
term "Moor" as stemming from the anxiety of the dominant Arab minority
about a possible alliance between African and Iberian Moors. While Arabs
may have had the sense that their rule throughout the région was unstable,
northwest Africans do not seem to have formulated a collective political
program against them.
The "Moors of Spain" did not defeat 'Abd al-Malik and did not offer his
kingdom to their allies. Yet, this is what the author claimed was their plan.
The generic "Moors" was definitely not an ethnographie concept. Instead, it

30 §75'- Wolf, Conquerors, 140; Lopez Pereira, 90.

31 §79: Wolf, Conquerors, 142; Lopez Pereira, 96.
32 §85: Wolf, Conquerors, 148; Lopez Pereira, 109-10.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
6o R. Rouighi / Studia Islámica 106 (2011 ) 49-76

was a political category, much like "the enemy" of military schools whose
plans are, by définition, always nefarious.33
As far as this source was concerned, the "Moors of Spain" pertained to a
specific political force in al-Andalus. In fact, every time they appear in the
text, the context is military or political. Obviously, since the chronicle is
primarily a record of important political events, this is hardly a surprise.
The association could not have been merely fortuitous, however, as the
"Moors of Spain" referred to a particular faction or party and did not refer
to members of a guild, a monastic order, or a heretical sect.
Similarly, the "Saracens of Spain" refers to a new group of Saracens. The
author recognized their ties with eastern Saracens, but viewed them as a
distinct group.34 As was explained above, the struggles that eventually led
to the victory of the Umayyads in al-Andalus were cast in terms of the lin
eages claimed by each of the contenders. In fact, the author's catégories
and distinctions are consonant with the struggles that led to the founda
tion of an independent Umayyad emirate in al-Andalus in 756. This politi
cal polarization in al-Andalus marked the author of The Chronicle of J54,
framed his vision of politics, and informed his recollection of the conquest
and its aftermath.

The sociopolitical specificity of the "Moors of Spain," their rébellion and

defeat at the hand of the Umayyads and their supporters, and their subsé
quent intégration into the Umayyad order as both subaltern and distinct,
illuminâtes the production of the generic Moors of northwest Africa. As
was seen, the politics in al-Andalus in the 730s and 740s accounts for the
emergence of two types of Moors: a concrete political force in al-Andalus,
and its obscure "natural" allies in the Maghrib. One was a social group, the
other a political construct.
The "Moors of Spain" did not appear in Latin and Greek sources prior to
the process of sociopolitical differentiation linked to the formation of the
Umayyad emirate in 756. While Arabie sources used by the anonymous
author are neither known nor extant, it is unlikely that they contained the
expression "barbar al-Andalus." If they did, the expression disappeared
without trace. It is more reasonable to believe that Arabie authors in

33 The description of the generic Moor recalls late antique conceptions dis
cussed by Moderan. Yet, as the appearance of the "Moors of Spain" demonstrates,
the older conceptions were adapted to the new political situation.
34 See for example §78: Wolf, Conquerors, 14t; Lopez Pereira, 92 and §82: Wolf,
Conquerors, 145; Lopez Pereira, 104.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 49-76 61

al-Andalus started employing the term "Barbar" to describe both the social
group and the political construct.
Once established, Andalusi conceptions of "Barbar," grounded in the
social and political realities of the península, circulated broadly in the
Maghrib and the Mashriq, and inevitably shaped writings about the Barbar.

Dynasties in the Maghrib

The victory of the Umayyads in al-Andalus was not the first time a dynasty
had claimed independence from al-Qayrawân. In the western Maghrib, the
Barghawäta (744/5-1058) were able to achieve their independence from the
governor of the Maghrib before the fall of Damascus.35 Another new polity
emerged in 788 near the ancient town of Volubilis in the western Maghrib.
The Idrlsids (788-959) rallied supporters and claimed legitimacy as descen
dants of the prophet Muhammad through his daughter Fätima and her
husband 'All b. Abi Tälib (d. 661). The basis of their power became obvious
when other 'Alid Arabs immigrated to the Idrlsid capital from al-Andalus
and Ifrïqiyâ and attempted to seize the reins of power. The locáis rebelled
immediately and made their military superiority felt.36 The Maghrib was
no Andalus.

Other northwest Africans seized power in the Maghrib under the banner
of Khârijism, a political idea that had been especially important in organiz
ing anti-Umayyad forces. The Khârijites, Sufrí and Ibâdî, insisted that legit
ímate rule did not belong to any particular group and that any pious Muslim
could assume the leadership of the Community. In the southern city of
Sijilmâsa, the Midrârids ruled for two centuries as Sufrí Khârijites. Further
north and east, the Ibâdïs (778-908), ruled from the city of Tähart over most
of the central Maghrib.
The early tenth-century chronicle of the Rustamid dynasty written by
Ibn al-Saghïr only uses the term "Barbar" once.37 Basing his information

35 See John Iskandar, "Devout Heretics: the Barghawata in Maghribi Historiog

raphy," The Journal ofNorth Ajrican Studies, vol. 12,1 (2007), 37-53.
36 See Ismä'il al-'Arabi, Dawlat al-adärisa: mulük ttlimsän wa fäs wa qurtuba,
(Beirut: Där al-Gharb al-Islämi, 1983).
37 Ibn al-Saghir, Akhbär al-A'imma al-Rustamiyln, Muhammad Näsir and
Ibrähim Bahäz, eds., (Beirut, Där al-Gharb al-Islämi, 1986), 52; and "Chronique
d'Ibn Saghir sur les Imams Rostemides de Tahert," Cahiers de Tunisie, 23 (1975), ed.
Mohamed Talbi, 331.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
62 R. Rouighi / Studia Islámica 106 (2011 ) 49-76

mostly on the oral reports of Ibâdï scholars, Ibn al-Saghir recounts a com
plex political history from 778 to 900 AH. The major groups he identifies are
the Arabs, the 'Ajam (i.e. Persians), and a number of MaghribI groups such
as the Nafüsa and Hawwära.

Although the author had many occasions to employ the term "Barbar" in
a generic sense, his single reference may be only the reflection of Ibn
al-Saghïr's usage, rather than that of his informants. Even so, it remains
impressive to see that more than a Century after the so-called "Barbar
revolts," a Maghribî author relying on his informants' catégories has very
little use for the term "Barbar." The degree of specificity achieved by giving
tribal names makes the generic superfluous. In addition, the absence of a
"Barbar" party in Rustamid politics and the prédominance of struggles
between identifiable tribes (Hawwära, Zanäta, Luwäta) help to explain Ibn
Saghlr's choices.38
The Aghlabids were the first independent dynasty to emerge in Ifrlqiyâ,
the région where Arab influence was greatest in the Maghrib. In 800, Ibn
al-Aghlab seized power in al-Qayrawân and soon after the 'Abbäsids in
Baghdad recognized him as a legitímate vassal. The Aghlabids (808-909)
were in a state of constant warfare with their Khärijite neighbors. They also
conquered and supported Settlements in Sicily. Aghlabid politics in the
island greatly resembled those of Umayyad al-Andalus. Importantly, as a
privileged source of information for the 'Abbâsid court, the elite of
al-Qayrawân played a key role in shaping eastern ideas about the Maghrib
and its peoples.
Ibn Saghlr's Ibädl contemporary Ibn Salläm (d. after 887) composed the
earliest extant text in Arabie written by a non-Arab northwest African.39 As
an Ibâdï scholar writing in Ifrîqiyâ, Ibn Salläm offered an Ibädl perspective
on Islamic tenets, legitímate rule, and Muslim learning. Naturally, as a
member of a group which opposed the Umayyad and 'Abbâsid dynasties,
he restated arguments showing the closeness of his ideas to the original
prophétie message. As a text, his Kitâb compiled written and oral sources.
Incidentally, ail the references to the Barbar are based on oral reports
collected by Ibn Salläm. Most are concentrated in only one out of the
twenty-one sections or chapters that constitute the book. When he copied

38 Assuming a very différent audience, al-Bukhâri distinguishes between the

Ibâdïs and the Barbar. See al-Bukhäri, Ta'rlkk al-rusuíwa al-mulük, 8:42.
39 Ibn Salläm al-Ibädl, Kitäblbn Satläm, EineIbaditisch-Magribinische Geschichte
des Islams aus dem g./g/Jahrhundert, Werner Schwartz and al-Shaykh Sälim Ibn
Ya'qüb eds., (Wiesbaden: Franz Steiner Verlag, 1986).

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica ro6 (2011 ) 49-76 63

from earlier written material, Ibn Sallâm did not refer to the Barbar. Henee,
it is reasonable to assume that "Barbar" was not yet a category hallowed by
tradition and usage. Most strikingly, Ibn Salläm reports a number of narra
tives (akhbär) which praise the Barbar (facLä'il al-barbar).40 Prominent
among these is a report about the prophet Muhammad in which the angel
Jibrll teils him that Islam will grow in the Maghrib and that the Barbar will
be its supporters.41 Another report contrasts the Arabs who fight for money
(1al-dlnärwa al-d.irh.am) and the Barbar who fight to establish the true faith.42
Overall, his narrative depicts the Barbar as devout Muslims. They were one
of the many non-Arab peoples who aeeepted Islam. Their participation in
"Islam," constitutes them as an ethnographie object clearly distinguishable
from the generic conceptions tied to the military conquest.43
In this light, Ibn Salläm's most explicit Statement about the political lan
guage of the Ibädls comes in the form of a report about their early leader
Abü al-Khattäb. When the Ibädls prepared to fight the Arab general Ibn
al-Ash'ath, a man called on "the people of Hawwära" (älHawwära) to rally.
Abü al-Khattäb immediately ordered the man flogged for using a tribal ral
lying call purportedly associated with the pre-Islamic period (Jâhiliya). He
urged his supporters to call on all good Muslims instead.44 Clearly, IbädT
politics sought to minimize tensions between those northwest Africans
who supported them by insisting on their new identity as Muslims. Yet,
despite their efforts, the Ibädls were not able to elimínate tribal catégories
and politics.
However laudatory his comments may be, Ibn Salläm's "Barbar" do not
explain political or social processes. Instead, his "Barbar" are a signal or tag
repeatedly appended to the names of groups or peoples such as Hawwära,
Zanäta, Nafüsa and Luwäta, as if to inform those who are ignorant of such
distinctions.45

40 Ibn Salläm, Kitäb, 121-25.

41 Ibn Salläm, Kitäb, 122.
42 Ibn Salläm, Kitäb, 124.
43 43. Although clearly beyond the scope of this essay, one can see that the
re-definition of Arabness, spurred by the migration of southerners into the Shäm
in the pre-Islamic period, continued in novel ways under the Umayyads and the
'Abbäsids. The discourse on universal intégration within "Islam" assumed an eth
nology, which was perforce shaped by events in both the Mashriq and Maghrib.
44 Ibn Salläm, Kitäb, 84.
45 The Arabs lack knowledge about the Barbar. The prophet Muhammad and
his caliph 'Umar b. al-Khattäb are described as inquiring about them. Ibn Salläm,
Kitäb, 122-3.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
64 R. Rouighi /Studia Islámica 106 (2011 ) 49-76

Furthermore, on more than one occasion, Ibn Sallâm translates Arabie

into "barbarlya," the language of the Barbar, which he obviously speaks. It
is not surprising to see someone who was educated in the culture and lan
guage of the Arabs and whose grandfather was an early supporter of the
Ibâdl Arabs use the language of the Arabs to describe his language in Arab
terms. This shows that elite northwest Africans had absorbed the catégo
ries of the Arab elite.46
Medieval Arabie authors did not distinguish between the sociopolitical
Andalusi "Barbar" and the generic and tended instead to confíate and jux
tapose them. Once they integrated the body of learned knowledge, how
ever, the narratives had a more tenuous and less obvious connection to
their own historicity.

The Barbar in the East

In comparison with the scarcity of extant sources from the Maghrib, ninth
century Arabie sources from the Mashriq are plenty. Writing in multiple
genres and disciplines, some of which were long established, and othersyet
in formation, authors combined narrative modes, stratégies, and procé
dures to compose multifaceted works.
The older idea that Barbar refers to people who live across the Red Sea
from Arabia persisted alongside the newly emerging knowledge about
Maghrib! Barbar. As I showed above, poems collected by Ibn Hishâm
(d. 834) mention the Red Sea "Barbar." In fact, they appear in an even
earlier source written by al-Waqidï (747-823). This 'Abbäsid judge of
Baghdad is known for his book about the early Muslim conquests entitled
Futùh al-Shäm (Conquest qfSyria).
Describing the Arab raids in the Egyptian région of al-Sa'îd (Upper
Egypt), al-Wâqidl enumerated the forces facing the Muslims during the
caliphate of'Umarb. al-Khattäb (d. 644).

In the Sa'Id there were Nüba, Barbar, Daylam, Saqäliba, Rûm, and Qibt;
and the Rûm were dominant.47

46 Ibn Salläm, Kitäb, 118.

47 Al-Wäqidl, Futüh al-Shäm, 'Abd al-Latîf'Abd al-Rahmän ed., 2 vols., (Beirut:
Dar al-Kutub al-'Ilmiya, 1997), 2:203. The "Barbar" do not appear in his Kitäb
aL-MaghäzL

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 49-76 65

Although it is not certain where these "Barbar" are located, their appear
ance in Company of such groups situâtes them in northeast Africa.48
Additional references in the text put them in the Company of Bajâwa,
Nüba, and Fallähln, and describe them with the Südän as people who use
éléphants in warfare.49
It is possible that these "Barbar" are "barbarians" in the Graeco-Roman
sense. However, such an assumption would require overcoming a few logi
cal difficulties. First, al-Wäqidl does not use this term to refer to people
anywhere but in southern Egypt. Second, their location so close to the
ancient Bapßocpia in the Red Sea area would have to be merely accidentai.
Notably, unlike pre-Islamic authors who refer to the people of the région
of Barbar (äl barbar), al-Wäqidl prefers the Barbar (al-barbar). Without
having to exaggerate the significance of a phenomenon that could be pecu
liar to this author, this shift allows the Barbar to be imagined as a distinct
people.
Sometimes an author used the category without the definite article (bar
bar), but then put it alongside other catégories that referred to known peo
ples. This is the case of the famous littérateur al-Jähiz (d. 869) who listed
"Barbar" among the Südän (people with dark skin), in Company of other
east Africans.50

Among the Südän are the Zanj, the Habasha, the Fazzän, Barbar, the
Qibt, the Nüba, Zaghäwa, Marw, the Sind, the Hind, the Qumär, the
Dablla, the Sin and Mäsln.51

Elsewhere, the historian al-Balädhurl (d. ca. 892) refers to a Störy about the
practices of the early Muslims about the imposition of the poll tax ( jizya)
on non-Muslims. It says: "the prophet Muhammad took the jizya from the
Zoroastrians of Hajar, [his caliph] 'Umar (d. 644) took it from the Zoroastri
ans of Färis (Persia), and [the caliph] 'Uthmän (d. 656) took it from Barbar

48 Al-Waqidî, Fatah al-Shäm, 2:212.

49 Al-Wâqidï, Futüh at-Shäm 2:214.
50 I prefer "people with dark skin" to Black people or Blacks because of the pon
derous connotations carried by such catégories in the modern period.
51 Al-Jähiz, Rasä'ilai-Jähiz, 'Abd al-Saläm Muhammad Härün ed., 4 vols., (Cairo:
Maktabat al-Khänji, 1964), 1:216. See also Ibn Khordädhbeh (d. 912), al-masälik wa'l
mamäUk, 155. One notes the appearance of the Chinese (Sin) among the Südän.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
66 R. Rouighi / Studia Islámica 106 (2011 ) 49-76

{barbar)."52 Since both Hajar and Färis are toponyms, it is reasonable to

infer that the Barbar meant here is the east African région.
Furthermore, knowledge about a defined entity of the Barbar seems to
have had an additional impact on the narratives. The verb "Barbara," which
means to speak unintelligibly in a language other than Arabie, came to be
associated with the Barbar. Al-Wäqidi uses the verb in its "common" mean
ing to describe the speech of non-Arab Byzantines.53 In this case, the "bar
barians" spoke Greek...
His near contemporary, the Andalusi 'Abd al-Malik b. Habib (d. 852 or
853), reported a very interesting variant. In the first exchange between God
and Moses, God spoke to Moses in the language of the Barbar {lisän al
barbar) and identified himself as the deity in that language, apparently
with little success.54 God then tried all the languages until Moses finally
was able to understand him. This story demonstrates that by the ninth Cen
tury, the Barbar had become an identifiable people with a respectable Bib
lical past. In addition, unlike the generic babble {barbara) described by
al-Wäqidl, Ibn Hablb's adaptation relies, at least partially, on the availabil
ity of empirical knowledge about the Barbar as a particular people with a
language of their own.
The association between the verb and the Barbar still provided the occa
sion for learned punning. Al-Balädhurl (d. ca. 892) reports one such
instance: "Ibn al-Kalbl [(d. ca. 819)] said: Ifnqush b. Qays b. Sayfi al-Himyarl
conquered Ifrlqiyä in the pre-Islamic period and so it was named after him.
He was the one who killed Jirjîr (Gregory) its king. He said about the
Baräbira (sing. Barbar): how prédominant is the babble of these people. So,
they were named Baräbira."55

52 al-Balâdhurï (d. 892), Kitâbfutüh al-buLdän, 3 vols., Salâh al-Dïn al-Munajjid

ed., (Cairo: Maktabat al-Nahda al-Misriya, 1956-7), 1:97-8.
53 Al-Wâqidï, Futüh al-Shäm, 1:5 and 1:19.
54 'Abd al-Malik b. Habîb, Kitäb al-Ta'rlj, ed. Jorge Aguade, (Madrid: Consejo
Superior de Investigaciones Científicas, 1991), §136,58.
55 Al-Balädhuri, Kitäbfutüh ai-buidän, 1:270. In fact, Ibn al-Kalbï's text is slightly
différent. It reads: "Ifriqush b. Qays b. Sayfi is the one who conquered (iftataha)
Ifriqiyä, and it was named after him. He also killed its king Jirjïr. And it is then
that the Barbar were named [for] he told them "How plentiful is your babble!"
Ibn al-Kalbï, Nasab Ma'ad wa al-Yaman al-Kablr, Nâjï Hasan ed., 2 vols., (Beirut:
Maktabat al-Nahda al-'Arabiya, 1988), 2: 548. See also al-Bukhäri (d. 870), Ta'rikh
ai-rusulwa al-mulük, 1:442.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 49-76 67

Obviously, this story projects the presence of the Yemenis in the Maghrib
to the period before the Arab conquest and the arrivai of the Syrians, with
clear political implications. The Yemeni "school" produced many such
stories, which seemingly attest their presence in the Maghrib in pre-Islamic
times.56

Medieval authors tended to be very sélective and favored reports whose

authority they trusted or supported. Consequently, they infused their
learned narratives with political concerns that are not always easily identi
fiable. Even if one cannot always decipher their signs, the early medieval
sources were not "confused" about the location or origin of the Barbar.
Instead, they incorporated heterogeneous material put together under dif
férent circumstances, for reasons that often remain obscure, and with vary
ing effects.

Western Barbar

The earliest extant historical narrative written in Arabie by a westerner is

Ibn Hablb's Kitäb ai-Ta'rïkh (Book of History). The author traveled to the
east in the 820s where he culled information about the past from historians
and storytellers, many of whom had ties to Egypt. His history, a universal
chronicle beginning before the création of the world, follows a loose chron
ological Organization. It includes entries on a number of topics, like "the
desire to accumulate wealth," which are not related to the succession of
events, rulers, and governors. Furthermore, Ibn Habib (d. 852/853) tended
to report information on the authority of eastern, rather than western
sources; understandably, there is little in the book that distinguishes it from
the eastern narratives.

As was already mentioned, the Barbar appear in Ibn Hablb's story of the
exchange between Moses and God. They are absent from the rest of his his
tory until the conquest of al-Andalus. Ibn Habib recounts that before the
Arabs crossed into Iberia, the governor Müsä b. Nusayr sent out men to the
Mediterranean coast with orders to capture Roman ships and maybe find
an older Roman man with some knowledge that could facilitate the con
quest. When they found just such a man, the Roman captive told them that
the Barbar would be the ones to conquer al-Andalus and that they would

56 See Ibn Qutayba (d. ca. 889), Kitäb al-Ma'ärif, Saroite Okacha ed., (Cairo:
Matba'at Dar al-Kutub, i960), 627-8.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
68 R. Rouighi / StucLia Islámica 106 (2011 ) 49-76

be Muslims.57 While it is possible that this indeed occurred, it is at least

reasonable to see it as a sanctification of the sense of inevitability and pre
ordained course of history.
Whatever the intention of its producer, this story clearly establishes the
"Islam" of the Barbar. In fact, this becomes the case with ail the subséquent
references to the Barbar, and over time develops into a recurring motif.58
This is not surprising since Ibn Habib does not mention the Barbar in his
brief description of the conquest of northwest Africa. His Barbar are in al
Andalus, where they play a politico-military role.59
In one of the many anecdotal stories, the author compares the martial
attitudes of the Barbar to those of various other groups. Among them, the
Barbar are most similar to the Arabs in terms of bravery. However, they are
also markedly différent, as they are reputedly most deceitful, disloyal, and
unlikely to respect any previously agreed pact. Alongside the Barbar, the
story refers to the people of al-Andalus (ahl al-Andalus), who are the Arabs
of al-Andalus, and are certainly not the Barbar.
Furthermore, mentioning the qualities that make Arabs and Barbar mir
ror images, simultaneously similar and différent, may have something to
do with politics doser to Ibn Hablb's time. Indeed, the author includes an
apocalyptic prédiction, according to which the Barbar will come to replace
the Umayyad rulers of al-Andalus. The "apocalyptic" Barbar rébellion trig
gers a series of others, which eventually lead to the end of times.60 The
Barbar are clearly dangerous and the idea of an overthrow of the Umayyads
is ail the more abhorrent because it seems plausible.

Possibly, someone added this last account to the chronicle after the death
of the author, at a time of serious challenge to Umayyad rule in al-Andalus.
Even so, it does not break from the formulaic représentation of the Barbar.
The negative sentiments of the Andalusi author toward a rival political fac
tion are probably to be expected, and henee they are less notable than the
generic and vague information he actually offers. As was demonstrated by
Mahmûd Makkl, however, Ibn Hablb's relative silence about the situation

57 Ibn Habib, §394,136.

58 See al-Bukhän, Ta'rlkh al-rusulwa al-mulük, 4:255. The argument against the
kidnapping of females was set in typically Islamic terms: *We found that [this prac
tice] is not condoned by book [Qur'än] or Sünna. And we are Muslims!"
59 Ibn Habib, 137.
60 Ibn Habib, §452,153.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 49-76 69

in the Maghrib and al-Andalus stems from his travel east, his intellectual
inclinations, and his Egyptian sources.61 In this regard, his Kitäb al-Ta'rikh
was the product of the same milieu that a génération later produced the
richer KitäbjutühMisr written by the Egyptian Ibn 'Abd al-Hakam (d. 871).
Ibn 'Abd al-Hakam begins his taie of the conquest of the Maghrib by
explaining that the Barbar migrated to the Maghrib from Palestine after the
defeat of their king Jälüt (Goliath) at the hands of the prophet Dâwûd
(David). Effectively, the story accounts for the géographie location of vari
ous Barbar in the Maghrib as well as the servility of the Afariq vis-a-vis the
Romans and anyone who takes over their land, which is a barely veiled ref
erence to the Arabs.62

The story anchors the narrative of the conquest of the Maghrib by iden
tifying the main characters that will play a role in the unfolding of the
events. As for character development, Ibn 'Abd al-Hakam's Barbar resem
ble those of Ibn Habib's. They attack, form alliances, rebel, or fall under the
command of a particular leader or authority.63 They are typecast as angry
adversarles, irrational rebels, and renegades who cannot be trusted, even if
they become Muslim. These Barbar are producís of tales by the vétérans of
the western conquests, as they play the part of the antagonists in the stories
of the heroic Arab generáis such as 'Uqba b. Näfi' and Müsä b. Nusayr.
On two occasions, Ibn 'Abd al-Hakam refers to specific Barbar subgroups:
the Luwäta and the Anbiya.64 The author gives no specific information
about them beyond their approximate location in the southern régions. It
is difficult to know from the text whether "Barbar" was added to their name

because of some sociocultural characteristic, because they were seen as an

adversarial politico-military force, or merely as a way of identifying them to
his audience as such. Ultimately, adding the Barbar tag onto the names of
northwest African tribes identifies them without explanation.

61 See Mahmud 'All Makki, "Egypt and the Origins of Arabie Spanish Historiog
raphy: a Contribution to the Study of the Earliest Sources for the History of Islamic
Spain," in The Formation of al-Andalus, Part 2: Language, Religion, Culture and the
Sciences, Maribel Fierro and Julio Samso eds., (Aldershot: Ashgate, 1998), 173-233.
62 Ibn 'Abd al-Hakam, 170.
63 Ibn 'Abd al-Hakam, 198 13; 199 3, 9,15; 200 2,14; 20112,17; 20416; 2051; 207 17;
20812; 21317; 214 3,14; 217 22; 218 2,4,5, 8 (unclear), 20; 2191 (copyist addition?), 13,
17; 220 5,11; 22210, 2231, 8; 225 2, 5.
64 Ibn 'Abd al-Hakam, 170,198.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
70 R. Rouigki / Studia Islámica 106 (2011 ) 49-76

Barbar Captives and Slaves

Many northwest Africans in the Maghrib integrated dominant Arab fami

lies as clients (mawält). As client, they came to accept the ways of those
Arabs who settled in the région. Naturally, Arabie sources tend to mention
the names of clients who were closer to the Arabs and their politics. How
ever, as the continued existence of powerful non-Arab tribes in the Maghrib
clearly shows, clientage was not the only type of relation that developed
between the Arabs and those they conquered.
Medieval Arabie sources agree that the Arab conquests also involved the
enslavement of a great number of northwest Africans. Ibn 'Abd al-Hakam
is very consistent in mentioning the number of captives and slaves the vic
torious Arabs obtained from the defeated. If one assumes that the numbers
given are accurate représentations, they would be staggering indeed. For
example, the author cites this report:

We were told by 'Abd al-Malik b. Maslama, and al-Layth b. Sa'd that

when Müsä b. Nusayr [d. 716-7] raided the Maghrib he sent his son
Marwän at the head of an army and [Marwân] captured 100,000. Then
he sent his nephew at the head of another army and he too captured
100,000. And so al-Layth was asked, "who are they?" and he answered:
the Barbar.65

From the inception of the conquests, Arab generáis such as 'Amr b. al-'Äs
required that the defeated pay the tax imposed on non-Muslims {jizya) by
selling their sons and daughters.66 The status of these slaves as non
Muslims was crucial since, at least in theory, no Muslim was to be enslaved.
Many of these "Barbar" were sent to Egypt and then onward to other
régions. References to Barbar slaves or servants abound. However, given
the graduai and slow imposition of Arab domination in the Maghrib and
the conversion of at least some Barbar, the "slave moment" would have
begun after the fall of Carthage in 698 and ended around the rebellions of
the 730s and 740s. Ibn 'Abd al-Hakam mentions that during the rule of the
Umayyad 'Umar b. 'Abd al-'AzIz (r. 717-20), "there remained [in Ifrlqiyâ]

65 Ibn 'Abd al-Hakam, 204. More conservatively, al-'Usfurï reports that the num
ber of captives was 20,000. See Khalïfa b. Khayyät al-'Usfurï (d. 854), Ta'rïkh Khallfa
b. Khayyät, Akram Diyâ' al'Umarî ed., (Beirut: Mu'assasat al-Risâla, 1977), 278-9.
66 Ibn 'Abd al-Hakam, 170. See also al-Balâdhurï, Kitíibfutüh al-buldän, 1:265.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011)43-76 71

not a single Barbar who had not become Muslim."67 One may see the events
of the 730s as a rébellion triggered by the illegal raiding by Arabs as well as
an indication of the importance of the slave trade to the dominant Arabs in
the Maghrib.68
One wonders about the impact of the slave trade between the Maghrib
and the Mashriq on the production of the chronicles, especially since these
most often include references to the enslavement of thousands of Barbar
after a battle, a seditious act, or betrayal of a legal agreement. Ail of these
were valid reasons to enslave them, because they demonstrated their true
character, belying their superficial conversion.69
In the Mashriq, the naming of slaves from the Maghrib involved the
addition of their "origin" (nisba). The preferred nisba seems to have been
"al-Barbarï." When used to refer to individuáis in the Maghrib, "al-Barbarf
was appended to a tribal nisba, functioning like a signal or tag to an eastern,
rather than western, audience. The différence is worth emphasizing here
because medieval authors tend to take it for granted.
Like modem labels, which identify commodities, "Barban" described the
slave, implied that he or she had certain qualifies, and influenced the price.
For instance, legal sources have recorded cases where the buyer of a
slave girl complained that he had purchased her with the understanding
that she was a Barbariya but then discovered that she was not. The opin
ions of the prominent jurists Mälik b. Anas (d. 795) and Sahnûn b. Sa'îd
(d. 854-5) demónstrate the existence of a fount of "common knowledge"
about the differentiation between slaves in the markets. They also show
that the Arabs who purchased slaves had a prédilection for the Barbar.70 In
describing the case, the author describes the theft and "illegal" enslavement
of freed Barbar in the Mashriq, perhaps because of their higher price they
commanded.

67 Ibn 'Abd al-Hakam, 213. See also al-Balädhurl (d. 892), Kitäbfutüh al-buldän,
1: 265; al-'Usfuri, Ta'rlkh, 323. 'Umar b. 'Abd al-'AzIz asked the Arabs who had a
female slave from the Luwäta to ask for her hand from her father or return her
to him according to the treaty signed by both sides. 'Umar b. 'Abd al-'Azïz's fiscal
policy eliminated différences between Arabs and non-Arabs.
68 The explicit link between slavery and rébellion is made by al-'Usfuri, Ta'rlkh,
380.
69 The use of "conversion" as a concept to characterize earlier political alliances
may itself be anachronistic.
70 Sahnûn b. Sa'ïd, al-Mudawwana al-kubrä Li al-Imäm Mälik b. Anas, (Beirut:
Där Sä dir, 19-?), 10:309.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
72 R. Rouighi / Studia Islámica 106 (20a) 49-76

Furthermore, these mercantile Barbar identify the origin of the com

modity but say little about the social or tribal affiliation of the individual.
No slave was described as belonging to the Zanâta or Luwäta, for instance.
By définition, slaves did not "have" tribes. Moreover, Arab men had an
incentive to have children from non-Arab slaves who lacked tribal affilia
tion because they had no kin who could claim part of the inheritance. Men
who attempted to guarantee that the property they accumulated would
remain with their kin group preferred slaves for those reasons.71 Obviously,
the majority of Arabs had fewer means. Their relations with northwestern
Africans, although less clearly known, did not necessarily exclude formai
marriage.
If slaves did not have tribal affiliations, at least some had marketable
skills, such as writing and accounting, which allowed them to gain social
prominence as individuáis. The influence of Hammäd al-Barbarî at the
court of the 'Abbäsid caliph Härün al-Rashld (r. 786-809) exemplifies this
phenomenon.72 The "Barbar!" origin, and thus special slave status (umm
walad), of the mother of the 'Abbäsid al-Mansür (r. 754-775) is also remem
bered in the chronicles. The social visibility of these famed slaves, or former
slaves, accounts for the repeated references to them as Barbar; henee, the
application of this label was not automatically a way to disparage someone.
Certainly, the slave merchants could only benefit from the existence of
such stellar exemplars that ultimately "proved" and confirmed the high
quality of Barbar slaves.
The memory of these individuáis' Barbar "origin" also points to the more
systematic forgetting of the lineages of tens of thousands of people living in
the Mashriq. In other words, this points to the absence of specifically "Bar
bar" political parties and "communities," such as armies or separate neigh
borhoods, in the east, contrary to the social conditions in al-Andalus—and
perhaps in Aghlabid Sicily and Ifrlqiyä as well. The différence is crucial. The
conversion of the Barbar slaves to Islam in the east led to their graduai
social "disappearance" there, whereas the politics in the west led to their
constitution as a differentiated social group.73

71 Sahnün b. Sa'ïd, al-MucLawwana, 10:309.

72 See Al-Ya'qübl, Ahmad b. Abi Ya'qüb, Tärlkh, 2 vols., M.Th. Houtsma ed.,
(Leiden: Brill, (1883) 1969), 2:498, 499; al-Balädhuri (d. 892), Kitäb futüh al-butdän,
1:58, 61.
73 Evidently, they did not completely disappear from learned knowledge.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011)49-76 73

In the same light, legal sources such as al-Muwatta by Sahnün b. Sa'Id

(d. 854) are richer than scholars have realized. They include information
about the legal différence between the Arabs who are known by their
fathers and tribes (nasab) and non-Arabs who are not. Such an idea takes
for granted the intégration of the non-Arabs into the Muslim polity and
their legal standing as clients (mawâlf) or former slaves. Examples of defa
mation suits demónstrate the legal différence between a Barbarï slave and
a Barbarï free Muslim who can press charges.74 They also show how the
preferential treatment of Arabs created distinctions between new Muslims
and thus conferred a new meaning upon the category Barbar.
The lack of tribal affiliation of the Barbar in the east could not have been
meaningless given that the issue of the lineages of the Arabs were of utmost
political importance given Umayyad and anti-Umayyad, and subsequently
'Abbâsid and anti-'Abbâsid politics. As explained previously, this period
witnessed coalitions of various groupings. Some of them were prevalently
Arab, or conversely, mostly anti-Arab in character. Others, however, were
based on alliances of Arabs and non-Arabs. The various ideological posi
tions they supported shaped the political scene and informed intellectual
production. These struggles, known collectively as "shu'übiya," also explain
why early Arabie authors have the tendency to inform their readers that
the Kutâma or Sanhàja were "Barbar." The addition of the moniker, or tag
ging, is part of shu'übiya politics. It played a crucial part in reinforcing the
Arab way of classifying peoples and tribes.
Furthermore, since the conversion of the slaves led to their immédiate
freedom, slave owners were not always eager to accept the conversion of
their slaves. Henee, while the Barbar were valued slaves, they were also dis
paraged lest they claim equal status. Al-Bukhârî (810-870) included in his
collection of Statements and deeds of the prophet Muhammad (hacLïth) a
number of unreliable and spurious ones. Among these, one captures the
anxieties of slave owners in the early Islamic period: "I heard the prophet
say that the faith of "al-Barbarî" does not go beyond his throat."75
Another aspect of the slave status of the Barbar in the Mashriq is the
attitude of the learned and pious men who wrote many of the sources still
extant. Another spurious hadlth states that the prophet said that "there are
seventy parts to foul; sixty nine belong to the Barbar, and one to the jinn

74 See §227 and §232 in vol. 15.

75 Al-Bukhäri, Kitäb al-ta'rïkh al-kabïr, (Haydarabad: Matba'at Jam'iyat Dä'irat
al-Ma'ärif al-Uthmäniya, 1941) 4 vols in 8, vol. 4, part 2, § 3389, 878.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
74 R- Rouighi / Studia Islámica 106 (2011 ) 49-76

and men."76 These types of Statements about the low moral qualities of the
Barbar, definitely espoused by Ibn 'Abd al-Hakam,77 are complementary to
the descriptions of Barbar slave girls as being the best entertainers. The
price of a single Barbar slave girl was reported to be 1,000 dinars.78 Slavery
and ideas about slaves among Arabs of property shaped the Arabs' attitudes
towards the Barbar, especially in this early period.
More generally, preferences among wealthy Arabs for slaves from par
ticular régions led to the formulation of a specific ideology tied to this activ
ity and this period. The slave trade stemming from the conquests brought
to the Mashriq individuáis and groups new to the région. Their intégration
into the society produced ideas about human différence according to which
the Arabs, as slave owners, belonged to the top of the social hierarchy. It is
important not to confuse this ideology associated with medieval slavery
with the modem concept of race. As I have also shown, the eventual inté
gration of the descendants of slaves and their melting away into the general
populace, as well as the drying up of the slave trade after the conversion
and independence of slave-producing régions, are some obvious différ
ences between modem and medieval practices. The incorporation of mod
em racial idéologies into the interprétation of medieval processes is
sufficiently problematic to warrant separate treatment.

Ancestors Twice Removed

As mentioned above, medieval authors fitted eastem and then western

Barbar into a discourse of origins which assumed a biblical vision of the
world.79 Following this established practice, Ibn 'Abd al-Hakam (d. 871)
recounts one of the most popular origin stories associated with the Barbar
by fitting his narrative into a biblical referential universe:

76 Ibn 'Abd-al-Hakam, 287.

77 In the first page of his Futuh, the author cites a Statement which describes the
world as a bird and the Maghrib as its ugliest part, the tail.
78 Ibn 'Abd al-Hakam, 220.
79 A more thorough examination is found in Maya Shatzmiller, "Le Mythe
d'origine Berbere: Aspects historiographiques et sociaux," Revue de l'Occident
Musulman et de la Mediterranee, N°. 35, (1983), 147. Cf. al-Bukhäri, Ta'rlkh al
rusul wa al-muiük, 1: 206, 207, 210, 604; Ibn Khordädhbeh (d. 912), al-masätik wa'l
mamälik, 80, 91.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
 R. Rouighi / Studia Islámica 106 (2011 ) 43-76 75

The Barbar were in Falastln (Palestine) and Jâlût (Goliath) was their
king. When the prophet Däwüd (David) killed the latter, the Barbar left
in direction of the Maghrib until they reached Lübiya and Marâqiya.80

The melting away of the distant past into mythological time and the con
comitant collapse of linear time typically mark these origin stories. More
importantly, medieval origin stories assume a biblical chronology, which
organizes time from Adam to the présent with Noah being an important
node, a second genetic moment. According to this view, Noah was a true
patriarch.
In asserting such a view, medieval authors could not, and did not, believe
that any particular people were "indigenous" to a particular géographie
région. The idea was not germane to their vision of the world and its past.
The earth was empty before the settlement of known peoples, and thus ail
of them were migrants who came to the land after the Biblical flood. As a
rule, nintheentury authors did not diverge from this general view: ail of
humanity was native to heaven, then after the fall came to possess guard
ianship of the earth. In other words, the question of origins was answered
fully by the story of the genesis. Interestingly, Ibn 'Abd al-Hakam's story
does not link the Barbar genealogically to Goliath. He was not their father
but their king. In fact, his account assumes that the Barbar exiles found
Romans and Afariq already settled in northwestern Africa.
As far as northwest Africans' own genealogies, the historian al-Ya'qübl
(d. 897) mentions that some groups of "Barbar" and "Afariqa" believed they
were descendents of Bar b. Nizär, while others believed they belonged to
Judhâm (Yemenis who migrated to the Shâm) and Lakhm (northern Arabs),
or Yemeni exiles.81 Al-Ya'qüb! also includes genealogies that tie the Barbar
to Noah, then enumerating known Barbar tribes, and explaining that each
had settled in that particular région of the Maghrib coming from the
east. While these genealogies are not too différent from those given by the
Arabs, they definitely agree that the Barbar had migrated to the région from
the east.

80 Ibn 'Abd al-Hakam, 170.

81 81. Al-Ya'qûbï, Tàrïkh, 1:215-6.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms
76 R. Rouighi / Studia Islámica 106 (2011) 49-76

Ancient and Indigenous Berbers

Catégories such as "the Berbers" and "the Arabs" are historical. Their pro
duction, maintenance, and reproduction occur under particular circum
stances. As circumstances change, so do these catégories. Their
représentation as everlasting and immutable is part of ideological procé
dures rooted in a présent.
In the early medieval period, the Arabs began a process of "Berberiza
tion" of northwest Africa, its peoples, and their pasts. Diverse social and
political conditions in al-Andalus, al-Maghrib, and the Mashriq, as well as
the évolution of literate writing in Arabie in the early medieval period,
account for the multiplicity and heterogeneity of the Barbar in Arabie
sources. The category was certainly not born fully formed. In fact, two cen
turies after the first raids, it was not stable, consistent, or uniform.
One of the ways early medieval authors secured the Berberization of
northwest Africans was through "tagging." This procédure classified the
Kutäma, Zanäta, and Sanhäja as Barbar. Interestingly, modem historians
have also used this technique to "produce" Berbers in ancient times. It is
easy enough to show that ancient authors never described the Libyans,
Numidians, or Mauretanians as Berbers, and that modem historians have
done so from the nineteenth Century and still do today. This should invite
an examination of the specific concems of modem historians.

While historians have focused on the similarity between Moors and Ber
bers, the medieval authors' récognition of the similarity between Arabs and
Barbar requires further attention. Indeed, the possibility that the Arabs
may have made the Barbar in their own image is replete with historio
graphie conséquences.
The analysis of the historical sources also shows that as late as the ninth
Century the idea that the Barbar were the original inhabitants of northwest
Africa was not firmly established. Henee, the idea was not the immédiate
product of the conquest, but rather of the gradual process of Berberization.
When modern scholars use concepts such as "indigenous" to refer to the
Berbers, they particípate in a related, yet différent and particularly modern
"Berberization" of northwest Africa and its peoples. The success of this pro
cess explains, in part, why historians continue to imagine that the Arabs
conquered a specific African people called the Berbers.

This content downloaded from

185.107.56.120 on Sun, 07 Apr 2024 13:11:50 +00:00
All use subject to https://about.jstor.org/terms