Introduction

PART I. MEDIEVAL ORIGINS
Chapter 1. Berberization and Its Origins
Chapter 2. Making Berbers

PART II. GENEOLOGY AND HOMELAND
Chapter 3. The Berber People
Chapter 4. The Maghrib and the Land of the Berbers

PART III. MODERN MEDIEVAL BERBERS
Chapter 5. Modern Origins
Chapter 6. Beacons, Guides, and Marked Paths

Conclusion

Notes
Bibliography
Index

* * * * *

Introduction

Berbers, Maghrib: the people, their country. Everybody knows that, and that is what everybody knows. But it has not always been the case. Before Muslim Arab conquerors began using the word barbar to refer to people who lived in what they called "the West" (al-maghrib), both people and region were known by a host of other names. In fact, before Berber and Maghrib, no one thought that the inhabitants of northwest Africa belonged together or that the entire landmass represented a single unit. The first time anyone thought that was in Arabic. Trying to understand this shift from one set of names to another, one map to another, a historian faces a series of challenges that can be separated into two general kinds. First, there are challenges arising from the handling of the sources, which tend to be late and not written to address such a question. Second, there are hurdles pertaining to the assumptions of modern historians. These include deeply held notions about the relation between collective identity (nation), country, religion, and language; and a centuries-long history of interpreting medieval sources in a way that reinforces these assumptions.

If this were not enough, as modern academics conferred on the Berbers characteristics of prenational groups like the Franks and the Goths, they also envisaged their reduced nationhood. For under French colonial domination, a modern Berber nation-state was simply not in the cards. After the Second World War, the reaction against the devastations of nationalism and racism did not extend to the category Berbers, which did not benefit from the critical energy of that reaction. Instead, the category remained mired in discussions of cultural heritage, victimized ethnic identity, and national aspirations. The national independence of Morocco and Algeria, but not Berberia, situated Berber identity both at the infranational level and as a counter nationalism but with a sense of Berber temporal precedence (native, original, etc.) and medieval Arabization and Islamization through a mixing with Arabs. These processes are reflected in the predominant place that anthropology, together with linguistics, occupies in the study of Berbers. Unpacking the entanglements created by modern relations and forms of knowing helps identify defining modalities of what it means to be Berber.

Together, these issues have combined to produce a peculiar consensus: the Berbers are the indigenous inhabitants of the Maghrib, their homeland. The basic idea is that even if the categories shifted after the seventh century, the people still had the same ancestors. For many reasons, however, this is not an acceptable position. No one would equate, say, Roman and Italian or Hun and Hungarian. Doing so would banish what historians identify as the stuff of history, and replace it with the stuff of ideology. But that is exactly what the early Arabic authors did. They displaced the old categories by adding Berber to them, creating a de facto equivalency among all of them: the Hawwāra became Hawwāra Berbers; the Zanāta, Zanāta Berbers; and all became Berbers. They also projected all these categories back into a remote past. They did not think the Berbers were indigenous, as moderns do, but they made them the descendants of Noah or tied them to some other ancient story. But for the historian, the phenomenon is still the same: at a certain point in time, under conditions that need to be ascertained, the Arabs began to populate their Maghrib with Berbers. I call this process Berberization, and it is the subject of this book.

Berberization was slow, but it eventually made associating the Berbers with the Maghrib seem natural. Even today, stating that the Berbers came to be thought of as the indigenous inhabitants of North Africa under specific historical conditions elicits immediate puzzlement. Other than academic historians and a few historically minded others who would find the historicity of a social category banal, most people might consider that whether one calls them Berbers or something else, the people were the same. And since their ancestors lived in the area for the longest time, they were indigenous. In the words of an anthropologist of Morocco, scratch a Moroccan, find a Berber. Perhaps, but thinking in this particular way is not natural, either. Instead, thinking historically about social categories—how they become ordinary, and how people use them to order their world—situates them in relation to both modern and premodern ideologies and scholarly crochets.

Defining Origins

Many studies on the Maghrib and its history begin with an attempt to define the Berbers as a way of introducing subject matter and cast of characters. These introductions usually include a discussion of the etymology of the word, its ties to the word barbarian and perhaps to the memory of the collapse of the tower of Babel. While offering a few anecdotes on the subject may satisfy the requirements of an introduction, a proper definition requires a degree of accuracy and coherence that has usually led those historians who have tried to define the Berbers to consult experts in related fields, such as anthropology, archaeology, and linguistics.

For historical reasons, French has been the language of the most serious attempts to define the Berbers. The best available synthesis in English is the one co-authored by medievalist Michael Brett and archaeologist Elizabeth Fentress, which takes into account the most important statements in the field and gives an accurate representation of the state of the question. As they endeavored to formulate a coherent definition of the Berbers, Brett and Fentress sought to clear a series of obstacles. Because of its quality, their definitional effort is a convenient way to introduce the subject to nonspecialists and give a sense of what this study intends to overcome.

Who are the Berbers and who counts as a Berber, according to Brett and Fentress?Just as the dialects are often mutually incomprehensible, so the people themselves are extremely heterogeneous: the existence of an ethnically unified "people" is no more demonstrable for the past than it is today. Indeed, there are a bewildering number of cultures, economies and physical characteristics. At best we can define Berbers as Mediterranean. In terms of their physical anthropology they are more closely related to Sicilians, Spaniards and Egyptians than to Nigerians, Saudi Arabians or Ethiopians: more precise characteristics are conspicuous by their absence, as a recent attempt at mapping a broad range of genetic traits has shown. We are thus immediately thrown into the problem of whom we are going to call a Berber and why.Immediately, Brett and Fentress encounter the problem of the ethnic heterogeneity of the Berbers, which appears as a problem only because they assume that the Berbers formed a unit of some sort. If not ethnographic unity, however, then perhaps physical anthropology—in other words, bodies and their appearance—could deliver a unity of "looks." It does not. After physical anthropology, Brett and Fentress make a foray into linguistics bringing into focus another basis for a definition of the Berbers: "The most common response [to the problem of whom one calls a Berber and why] is linguistic: Berbers are defined as people speaking Berber languages.. . . Indeed, one of the things that sets the Berbers apart is their language.. . .This was often commented on in the past, and a common myth links the odd-sounding language to the name 'Berber.'. . . The Berber dialects are part of the language group, the Afro-Asiatic, which comprises the Semitic languages and Ancient Egyptian." Linguists use formal properties of living languages such as Sīwī and Arabic to classify them within language families like Berber and Semitic. The study of a number of related languages allows linguists to reconstruct the features of the parent language. So, although proto-Berber and proto-Semitic are not extant or attested, it is still possible for linguists to know enough to distinguish between them, even if, in the case of these two families, they share a great many features because both split from the same parent language known as Afroasiatic. Understandably, dating the differentiation of undocumented languages and situating their bifurcation geographically is complicated and involves a lot of guessing. Yet, there is a great deal of good science behind it. When it comes to proto-Berber, the consensus is that it split from northern Afroasiatic somewhere in eastern Africa and then spread westward from there. There is less of a consensus about the date of that event, or events, but it varies from around 9,000 to only 3,000 years ago—a staggering range. Even as they work to reach more precise estimations, however, for linguists the question of the origins of Berber is largely settled. Like Arabic and Punic, Berber came from the East, just earlier than they did.

Brett and Fentress repeat a statement that is very important among specialists: "What sets the Berbers apart is their language." That is a good basis for deciding whether an individual or group is Berber:The ability to speak a Berber language gives us an objective basis for asserting that a given individual is Berber . . . but if we restrict ourselves to a linguistic definition of Berbers when discussing their history there will be few groups we can discuss with certainty. A cultural definition appears more promising, but when applied to the past becomes unsatisfactory. Unfortunately, this is a common procedure: perceptions of Berber culture derived from modern anthropology are often casually back-projected to antiquity. Worse, they are then used to justify a judgment that Berbers were culturally immobile.Brett and Fentress are correct: until the seventeenth century, sources in Berber languages are relatively rare, fragmented, and not evenly distributed. They do not explain why they believe that the extremely heterogeneous Berbers and their bewildering number of cultures, economies, and physical characteristics have a single history. But they are right about the circularity involved in projecting modern representations into the distant past and then using them as evidence of cultural stasis. Surprisingly, the conclusion Brett and Fentress draw from this sensible observation takes them in an entirely different direction: "The least unsatisfactory solution seems to be to use the term 'Berber' in the broader sense of those groups who were perceived to be indigenous North Africans, both in antiquity and in the middle ages, as well as anyone who is still perceived that way today." More significant than the lack of sources to shed light on what people might have perceived in the past, let alone whether they could even have perceived someone to be an indigenous North African, the notion of indigeneity allows Brett and Fentress to tag those human beings for whom there are only archaeological artifacts as Berber or proto-Berber—not to be confused with the proto-Berber of linguists.

The notion that all prehistoric human settlements found in North Africa are related to the Berbers is not universally accepted, however. In her excellent presentation of the state of archaeological knowledge on the subject, Malika Hachid argues that Capsian and Mechtoid civilizations combined to form proto-Berbers at a particular time, between 11,000 and 10,000 BP. This is how she explains it: "If we have somewhat insisted on Capsian portable art and then on parietal art from the Saharan Atlas it is because we consider Capsian Protomediterraneans to be the artisans of Berber identity and culture to which the Mechtoids contributed as they integrated. When it comes to their language, if the Capsians brought with them the rudiments of the Berber language, they could not but absorb some aspects of the language of the Mechtoids." For Hachid, the Capsians "brought with them" the foundations of the Berber language from the East, which makes it, but not them, not indigenous. By absorbing some of the language of the Mechtoids, the Capsian Protomediterraneans indigenized Berber—at least in part. The total absence of evidence of Mechtoid or Capsian languages is not critical because Hachid's Berberness (Berberité) is as it turns out tied to art. In spite of Hachid's timorous statements about the lack of "a perfect homology between human type and culture," she uses categories such as "robust negroid" and "fine negroid" to discuss the geographic, ethnic, but not racial, origins of people who might have combined to form the proto-Berbers. More than how she arrived at her conclusion, however, the idea that proto-Berbers emerged only 10,000 years or so ago, or maybe only 7,000 years ago in the Sahara, leaves us with a very long time of non-Berber human presence in the area from the Atlantic to the Nile and from the Mediterranean to the Sahara. Likewise, without a linguistic definition, there is no reason why the proto-Berbers could not be proto-Algerians or proto-Maghribīs.

But obviously, as Brett and Fentress' definition illustrates, not everyone is willing to give up on the linguistic factor, "one of the things that set the Berbers apart." Naturally, the appearance of proto-Berbers is not innocuous. It is tied to the question of indigeneity, which has not been the friend of historicizing, at least when it comes to preventing anachronism. In any case, from the point of view of indigeneity, the word Berber is problematic: "Of course, even the use of the name 'Berber' is somewhat arbitrary: it is of external origin, and certainly not a Berber word.. . . The word for Berber today is either 'Tamazight' or 'Imazighen', the first referring to their language, the second to the people who use it." What is significant here, more than what category to use, is that Brett and Fentress deem the word Berber unsuitable, not because it lumps together a multitude of groups and occults the temporal specificity of documented collective categories and not because there is anything wrong with the linguistics behind it, but purely because it is an exonym.

They contrast the foreignness of the category with the indigeneity of the people. This too is critical. For behind the act of defining is the discourse on the indigenous origins of the Berbers. But since each discipline produces its own timeline for the "emergence of the Berbers," definitions that incorporate all these timelines have a hard time reconciling them. If there is a consensus among archaeologists, linguists, and biologists, however, it is for dating the origin of Berber long before "the term [was] first recorded in Arab authors." What should stand out here is the reliance on philology (etymology), physical anthropology and archaeology, cultural anthropology, and linguistics—in short, everything but history—for a definition of the Berbers.

Since Brett and Fentress mention genetics and since that subfield of biology has been completely transformed following the sequencing of the human genome in 2000, it may be useful to highlight the contribution of biology to the Berberization of the remote past. When it comes to biology, however, the word origin does not always refer to the same event. On the one hand, there are the origins of all life-forms on earth, which go back millions of years. As complex living organisms, the Berbers have their origins in early amino acids and other life-supporting molecules. The Berbers are also great apes and therefore share with all humans an original moment of distinction from other hominids. In this respect, the origins of the Berbers are the same as those of New Yorkers and Indonesians.

In another register, the question of origins refers back to biological differentiation between human populations in time. For this, specialists use a naturally occurring mutation that passes on to offspring to date the emergence of an ethnicity. In other words, the first bearer of the marker is the common forebear of an ethnos, a statistically defined population. As often happens, however, the mutation that best identifies an ethnic group through the male line (Y chromosome) does not quite match the chronological information carried by the female line (mtDNA), and the common male ancestor of a given group lived sometimes hundreds of years and thousands of miles apart from its common female ancestor. For instance, the most common male marker among modern Berbers (E-M81) goes back 5,600 years, whereas some mtDNA lines go back 50,000 years—close to the time human language is thought to have first emerged.

Members of a population such as the Berbers carry a multiplicity of markers, some of which are much older than any evidence of modern human beings in northwest Africa. In other words, some of the ancestors of the Berbers must have come from somewhere else, something that hardly makes the Berbers special. Moreover, if one identifies the Berbers not with those markers prevalent among current populations, as is common practice, but rather with those mutations that took place exclusively in northwest Africa, only a fraction of today's Berbers would qualify. But when they collect data from contemporary speakers of Berber languages and use them to produce Berber ancestors who lived thousands of years ago, biologists contribute to populating that past with Berbers. Until they test ancient and even fossil DNA, Berber DNA remains tied to modern populations who speak any one of the many Berber languages. Once we have Numidian DNA, we will have a different kind of problem on our hands. Until then, it is fair to say that there is still a difference between being biologically related to someone and belonging to the same social group—as Americans know all too well.

Parallel to the effort to cast the Berbers as indigenous is the drive to show that they have a privileged relation with the land, their homeland (Heimat), and that they therefore have a more legitimate claim to it than anyone else did in the past or does in the present. But rather than try to set up indigeneity tests based on recent and even fossil DNA, attention to Berberization brings into focus the earliest time it was even possible to think that any group belonged in "North Africa." Unlike the formation of the protolanguage, the timing of a frequently occurring mutation, or the types of snails prehistoric groups seem to have really enjoyed, that is an event that historians can illuminate.

Historical Origins

No one was a Berber in northwest Africa before the seventh century, and that is when this study must begin. Then again, since the earliest extant Arabic sources are from a little more than a hundred years later, that starting point is somewhat truncated. To be more precise then, this study begins with the earliest references to Berbers in Arabic sources. What is striking about early Arabic usage is that the word Berber did not refer to anything like an ethnic group, not in the late antique Greek and Latin sense of what "ethnic" meant. In fact, it did not even mean anything like a "people," or at least not exclusively so. This means that more than a century after the first raids of the 640s, the transformation of various northwest African groups into Berbers, their Berberization, was incomplete. However, since the groups that came to be called Berber in Arabic had not formed a unified entity prior to the conquests, this is not very surprising. Had the Arabs found a single kingdom that ruled over what they called the Maghrib, they could have called its people Berbers and that would have been that. But that is simply not what happened. A little like the Indians of the Americas, most Berbers did not even know they were being Berberized, at least not for a while. Moreover, had the Arabs conquered only the region immediately west of Egypt, their Berbers might have been geographically limited to that area. They might have eventually called people in the western Maghrib Berbers too, but that is not necessary. The counterfactual highlights the importance of the conquests, an importance heightened by the circumstance that the earliest Arabic sources that mention the Berbers are narratives of these very conquests.

Not just chronologically, but also conceptually, Berberization begins with the conquests. The shift of perspective solves the problem of anachronism, while also making better sense of what took place. Imagining the conquests as triggering the Berberization of various groups allows us to resolve many of the apparent contradictions found in the sources and supports coherent historical explanations that connect events with their representation in the sources. The same approach can then be extended to later moments when the discourse on the Berbers was transformed, augmented, and adapted. Of course, treating the category as if it remained stable throughout the medieval period is also anachronistic. Methodologically, the study of the conquests can serve as a model of how to handle later instances of Berberization. In order to highlight this point, the book will describe multiple sites of Berberization and thus multiple historical origins. For without the reproduction of Berbers, they would have simply disappeared like the Numidians and the Getulians.

Framing the historical problem in these terms draws attention to the formation of an imperial knowledge in Arabic and to its circulation across a vast area. It also points to the eventual emergence of the Maghrib as the primary, if never lone, center of Berberization. In other words, our multiple sites were chronologically and geographically distinct. That, in turn, suggests that the historical record may not include all instances of Berberization since documentation on some of these sites may not exist. Yes, this poses a serious challenge to this attempt to understand Berberization, but historicizing begins with recognizing and accepting the limits of the knowable.

This book is not a history of the Berbers. The focus here is not on what happened to the Berbers, but rather on how it became possible to think that something happened to Berbers in the first place. The only claim this book makes is that the process was and remains historical. But as the discussion of how Berbers have been defined shows, a study of the making of Berbers faces multiple hurdles. By assuming the existence of Berbers in ancient times and by using the category as if what it signified remained unchanged over the entire medieval period, modern historians have engaged in a Berberization of their own. In fact, this modern Berberization has set the terms of historical research about a whole set of subjects, the conquests of the seventh century being only one. Berberization has even informed the prevailing periodization. These are not marginal considerations. And since the modern historiography of the medieval Maghrib has been entangled in the experience of French colonial domination and its aftermath, difficulties abound. Insisting on the historicity of social categories like Berber allows us to sidestep issues tied to the impact of colonialism and colonial historiography. However, just as it does not try to blame or shame medieval elites for exploiting their social subordinates or medieval intellectuals for representing the Berbers in ways that fit dominant ideologies, the book could not possibly fault modern intellectuals for doing the same. That is not the point here, even if it is conceivable that someone might misunderstand that.

The Mother of All Texts

At the heart of the functioning of Berberization in modern scholarship lies the Khaldunization of knowledge about the Maghrib, which began in the nineteenth century. Ibn Khaldūn's work is not just fundamental to the constitution of modern historical knowledge about the medieval Maghrib, but is part of the very matrix that has generated that knowledge itself. Without Ibn Khaldūn, there is simply no medieval Maghrib. He stands at the end of the period and is, given his importance, the last medieval author. The Khaldunization of the field of knowledge has authorized, validated, and oriented scholarship, determined its questions, weighed on its explanations and interpretations, and even inspired its speculations.

The Ibn Khaldūn in question here is not the fourteenth-century author of the Kitāb al-'ibar (Book of Examples) and its famous introduction (al-Muqaddima). Instead, he is the author of the Histoire des Berbères (1852-56), de Slane's French translation that gave language and structure to modern studies on the medieval Maghrib. Through repetition, duplication, and multiple variations on set themes, historians constituted a Khaldunian historical field, and as they did so, they established the textual basis for projecting the Berbers into the pre-Islamic past, which has precluded the identification, let alone study, of Berberization. On the one hand, the language of de Slane's translation made it harder to imagine that the Arabic category could have emerged in time or evolved over time. More importantly, the quality of Ibn Khaldūn's thinking and the breadth of his work conferred on the text an immediate authoritative character, especially because other medieval texts were either unknown or not easily accessible, and many of those who jumped on the Ibn Khaldūn bandwagon preferred French to the original Arabic. By the time the first serious historical studies on the medieval Maghrib began to appear, there was already consensus on the importance of Ibn Khaldūn. This explains why the framing of these studies, the questions they raised, and the periodization they established can all be attributed to this one text.

While it is critical to situate modern Ibn Khaldūn in modern Berberization, it is also necessary to remember that the work of the historian Ibn Khaldūn crystallized the extent of Berberization of discourse in the fourteenth century. Through Ibn Khaldūn, nevertheless, medieval Berberization comes to shed light on the constitution of a modern historical field and the specific conceptual entanglements it has created for historians, and—although not the focus of this book—for nonhistorians too.

Organization

This book is divided into three parts, each with two chapters. The first part, entitled "Medieval Origins," examines Arabic sources on the Berbers and relates them to political and other developments that illuminate them. Chapter 1 treats the question of origins as one about chronology and tries to ascertain the historical conditions that best illuminate the earliest references to the Berbers in Arabic. Since there was no preexisting people whom the Arabs simply called by a different name, the word Berber carried a range of meanings associated with specific social activities and institutions, chief among which was the military. The chapter contrasts the political situation in al-Andalus and the Maghrib to establish a shift in the usage of the category. Chapter 2 attempts to enshrine the historicity of the category by examining multiple sites of Berberization, from al-Andalus to the Maghrib and Egypt. The chapter demonstrates that the category did change, and that it was not the same everywhere.

In the second part of this book, the focus shifts to an analysis of the idea that the Berbers were a people and that the Maghrib was their homeland. Chapter 3 explores how medieval authors envisaged the Berbers by analyzing the functioning of genealogy ('ilm al-nasab or 'ilm al-ansāb), paying close attention to their classificatory schemes and categories. But rather than consider the entire Arabic archive, the chapter examines a selection of documents that shed light on the organization and content of Ibn Khaldūn's work. Since Ibn Khaldūn's text comes to have such an immense place in modern historiography, the chapter serves also as an introduction to his medieval work. In the same fashion, Chapter 4 tests the idea that Arabic sources have always thought of the Maghrib as the country of the Berbers (bilād al-barbar). It too leads to Ibn Khaldūn and to his ideas on the subject. Looking to explain the circumstances that allowed medieval authors to conceive of the existence of the Berbers and the Maghrib in the remote past, these two chapters identify elements in the constitution of modern historical thinking on these same topics, with Ibn Khaldūn serving as the link between the two.

In Part III, modern Berberization comes into focus, through a consideration of select sites of modern Berberization. Given the importance of French colonial rule in Algeria, the chapters are concerned more narrowly with Algerian developments to identify critical elements in the modern making of the Berbers and to trace their evolution over time. Chapter 5 argues that the publication of de Slane's translation of Ibn Khaldūn was a major event in the formation of modern knowledge on the Berbers. Chapter 6 analyzes the historiography of the medieval Maghrib in the last century in order to explain why and how Berberization came to be hidden.