From herd to horde : the emergence of nomadic pastoralist armed conflict : a Tuareg case study

(1)

“We drink their milk, we eat their meat, we use their skin, we trade them. When the animals die, the Tuareg dies” - Gwin, 2011

Jules Goslinga - 10685804 - University of Amsterdam (UvA) Political Science, International Relations track:

Civil wars, state-building and displacement research project

From Herd to Horde: The Emergence of Nomadic Pastoralist Armed Conflict - A Tuareg Case Study

Supervisor: Dr. Abbey A. Steele - University of Amsterdam (UvA)

Second reader: Dr. Darshan V. Vigneswaran - University of Amsterdam (UvA) 22 June, 2018

(2)

Acknowledgements

Throughout the process of writing this thesis, I have had the support of a few people whom I would like to thank. First of all, I am grateful for the supervision of Dr. Abbey Steele, who was of great help in guiding my research and, together with Chris Wignall, provided much appreciated feedback. I would have like to thank all family and friends who helped me stay motivated during a period of independent and largely solitary research.

(3)

Abstract

As the process of globalisation continues, pastoral nomads are increasingly experiencing changing conditions, with all due consequences for their lifestyle. This investigation aims to observe if and how restrictions on nomidisation and pastoralism contribute to the emergence of armed conflict. Through secondary case study research, the Tuareg of the central Sahara are investigated to determine the role and process of these restrictions in the four periods of rebellion that have occurred since the withdrawal from the region of the French Empire circa 1960 and until 2013. The results, focussing on Mali and Niger, suggest that although the restrictions on nomadisation are influential in the outbreak of violence, there are additional causes. Therefore, this study adds to the literature some specificity in the niche field of nomadic violence, but does not develop comprehensive theory.

(4)

Figures and tables

❏ Figures

❏ Figure 1: Gurr’s relative deprivation graphs 14

❏ Figure 2: Tuareg confederacies map 35

❏ Figure 3: Historic Tuareg territory map 37

❏ Figure 4: Saharan trade route map 38

❏ Figure 5: Tuareg conflict timeline 39

❏ Figure 6: Malian ethnic distribution map 46

❏ Figure 7: Second rebellion organisations chart 49

❏ Figure 8 Government changes timeline 63

❏ Figure 9: Human-induced restrictions timeline 64 ❏ Figure 10: Sahelian precipitation anomalies timeline 65

❏ Figure 11: Causal mechanisms diagram 66

❏ Tables

❏ Table 1: Observable implications 28

❏ Table 2: Tuareg confederacies 35

(7)

Introduction

Amongst the most epic and iconic cases of pastoral nomads turned warriors, is that of the Mongol horde. Prior to Genghis Khan's transformation of his tribal following, priority lay with finding pasture and exploiting the produce supplied by the herd (Biran, 2004). This shift, which occurred more than 700 years ago, has been studied thoroughly and demonstrates that nomads - communities which are generally characterised by their relative isolation from their respective neighbouring societies - can under certain circumstances go to war, a phenomenon that happens to this day. Some argue that to a notable extent, among the “contributing causes for the frequent and widespread incidence of various forms and degrees of warfare [is the] factor of the character and nature of the nomad himself” (Ekvall, 1964, 1125). Through the habituation to danger from early childhood amongst the herd, nomad children are brought up “to endure, to take chances, and to react to all emergencies with immediate and explosive effort” (_Ibid.

, p. 1126).

Globally, it appears that there are different instances of nomadic peoples groups becoming involved in violent conflict. From a distance, one might not expect much interaction between nomads and others, let alone armed conflict, but various examples, which will be discussed, indicate that indeed their mobility (and corresponding marginal position within a larger society) does not exempt them from conflict. In his book ‘Nomads and the outside world’ A.M. Khazanov writes that “it is important to keep in mind that [...] pastoral nomads, were culturally and ideologically dependent upon sedentary societies, just as they depended upon them in economic respects” (Khazanov, 1994, p. XXXI-XXXII). Moreover, “anthropological research shows that seldom, if ever, have pastoral nomadic societies been able to prosper in isolation from other economies, in particular agriculture-based ones” (Di Cosmo, 2002, p. 168).

With the steady disappearance of nomadic lifestyles around the world (Annamoradnejad & Lotfi, 2010; Keenan, 2003b), comes a phase of transition for many ethnic groups, with some asserting that “nomadic pastoralism may have been productive thirty or forty years ago but declined due to increased drought [in many

(8)

parts of Africa] and it was now necessary to abandon pastoral mobility for settled livestock and crop farming” (Frantkin, 2014, p. 101). Facing globalisation, expanding spheres of outside influence, the advent of the nation-state, as well as opportunities never seen before, their “social and political position has weakened due to the penetration of the modernisation process in [many] aspects of [...] society” (Annamoradnejad & Lotfi, 2010, p. 335). Although with the colliding of these worlds violence is not always a necessary occurrence nomadic peoples do have a global reputation of being inclined towards violence (Ekvall, 1964).

This investigation aims to provide new insights on the relevance of such changing circumstances, as experienced by nomadic peoples, with regards to the outbreak of armed conflict.

First, on the basis of case study research, an assessment is made of the range of factors affecting nomadisation, and their ability to contribute to causing the outbreak of intergroup violence. In other words, the investigation starts by examining the causal effects. In the subsequent phase, the causal mechanisms of these causal effects are analysed. What this entails is a more indepth look into the processes, asking the question of _{how those causal effects that are determined to be influential} in the case study actually lead to armed conflict.

Definitions

The definition of ‘nomadic peoples’ that is used over the course of this investigation will relate primarily to pastoralists. As such, an appropriate description of the term is taken from Crawford and Leonard (2002, p. 1-2), who write: “nomadic pastoralism refers to populations specialising in animal herding, which requires periodic movement for purposes of grazing”.

“Pastoralism is above all a flexible subsistence strategy, involving opportunistic food production and foraging in addition to livestock exploitation for meat, milk and blood”, indicating often varying means of production (Crawford & Leonard, 2002, p. 99). As will be displayed later on in the case study, nomadic pastoralism is often not sufficient for obtaining all necessary resources a population requires. Therefore, trade for instance, is a common way to complement it. In other

(9)

words, pastoralism is the main form of resource production, but rarely the only one within a society.

A brief look at the etymology of the term ‘pastoral nomad’ reveals the core elements of what it means to identify a person as such. The word ‘nomad’ comes “from Greek_{nomas [meaning] ‘roaming, roving, wandering’ (to find pastures for flocks} or herds), [and is] related to _nomos

[meaning] ‘pasture, pasturage, grazing’” (Harper,

2018a). The adjective ‘pastoral’ comes “from Latin _{pastoralis [meaning] ‘of herdsmen,} of shepherds’” (Harper, 2018b). Evidently, both stem from a similar root and their synthesis clearly points at the image of the roaming herder, leading his livestock to fertile grazing lands as a livelihood.

Following from the above, _{nomadisation can be defined as the ability to lead a} lifestyle in correspondence with the nomadic identity. In reality however, the term can be rather vague. The subsequent citation raises a point that should be mentioned:

“_{In any given situation in any region, the nomadic pastoralists,} who ever lived on the peripheries, because of their basic mobility and the considerable evasive capability that mobility conferred, had even greater local autonomy and independence than that possessed by other segments of the population

” (Ekvall, 1964, p. 1122).

In other words, Ekvall says that all nomadic pastoralists, due to their mobility, have greater autonomy than other populations. Therefore, arguably, nomadisation is inextricably linked to self-determination and autonomy. As this understanding of the term could obstruct the drawing of any meaningful conclusions and provoke excessive confusion, autonomy in itself shall not be considered as inevitable to nomadism.

Instead, a more practical definition, provided by Dyson-Hudson and Dyson-Hudson is used. It states that pastoral nomads are “those people who are principally dependent for their subsistence on livestock, and for whom spatial mobility is regularly employed as a survival strategy” (Dyson-Hudson & Dyson-Hudson, 1980, p. 15). These two elements are key to the subsequent investigation.

(10)

Research questions

Prior to the primary research question, it is tested if the assumption holds that nomadic peoples are in fact involved in conflict with a certain regularity, and not just sporadically. In other words, the descriptive question asks; do nomadic people become involved in armed conflict? To answer this, descriptive statistics are presented, aimed at supporting the notion that deadly, targeted violence does indeed occur in the selected Tuareg case study. This serves to evaluate and support the relevance of the actual research question.

Given that the results to the above question are convincing of the notion that nomadic conflict is real and therefore an area worthy of studying, one can begin to consider its origins. Hence, the general research question is introduced. Derived from the interest in abovementioned Mongol transformation, it reads; _{why and how do}

nomadic peoples become involved in armed conflict?

The case study which is chosen to be studied in further detail is that of the conflict involving the Tuareg in the period following the decolonisation of former French West African colonial empire - containing practically all of the historical region of Tuareg influence - until the present day. It is widely known that violence is part of their recent history, yet the reasons explaining this are far less understood. Selecting this case has several benefits and difficulties that will be elaborated on at a further stage. Now, for the sake of clarity, it is deemed helpful to reiterate the research question once again with particular regard to this case study. As such, it is asked;

why and how have the Tuareg become involved in armed conflict with state governments and others in the different violent periods since the decolonisation?

Both questions are formulated to help guide the investigation that follows, aiming to provide a coherent answer.

Approaching the research question

As mentioned above, the approach to the research question is primarily concerned with utilising the Tuareg case study to develop inferences about the underlying causes for conflict involving nomadic peoples. As such, the initial focus

(11)

lies with answering the case-specific research question. Provided that a valid and reliable response to the case research question is produced, attention may shift towards the resolution of the general research question. I will pay attention to differences and similarities of the Tuareg vis-á-vis the larger nomadic world and deriving the implications from the case which may be applied more generally. Nevertheless, if deemed helpful to the flow and structure, the general RQ may be referred to and commented on intermittently, whilst discussing the case study.

Efforts are made to structure the answers of both RQs in two stages, set apart by the two types of causal inference that are to be made (NB: descriptive inferences are only made in gathering evidence that pastoral nomads, and the Tuareg in particular, are in fact in conflict, prior to exploring causal relationships). The first stage focuses on making observations in response to the ‘why’ component of the RQ; i.e. what the causal _{effects - or explanatory factors - are that significantly contribute to} the emergence of conflict in relations to nomadic people groups.

Having discussed the ‘why’ question, I consider it incumbent to attempt to answer _{how these causal effects prompt armed conflict. In doing so, causal} inferences are to be made in order to develop a plausible and convincing causal mechanism - defined as “independent stable factors that under certain conditions link causes to effects” (George & Bennett, 2005, p. 8) -, illuminating the process of how the causal effects materialise and steer the Tuareg and other nomadic ethnic groups towards violence.

The method of data collection that is employed in this investigation is secondary research, consulting existing sources, such as academic studies, government or NGO reports, news articles, etc. There are both advantages and disadvantages to this choice, the rationale of which is further elaborated upon in the research ‘design section’.

Finally, to comment on the importance of the subject and its further exploration, nomadic people are generally a group that is relatively little studied, often living on or beyond the margins of (sedentary) society. In fact, nomadic ethnicities are often misrepresented by outsiders, and “the idea that nomads are disinclined to engage in any activity other than contemplative herding, caravan trade, and warfare,

(12)

are hard to root out, despite the many detailed ethnographies depicting pastoral life as one of constant toil and labour” (Lecocq, 2010, p. 129).

Seemingly, existing literature rarely specifies with particular interest how general theories and ideas in the field of international relations relate to nomadic peoples.

Although the Tuareg and other nomadic tribes have previously been studied to a considerable extent - in fact, even their relation with violence has been observed - there is no general unifying idea of how nomadic peoples become involved in armed conflict. In other words, despite earlier conclusions with regards to the reasons for war for individual nomadic ethnic groups, those findings have not resulted in inductive reasoning about generalisable conclusions.

As such, this paper aims to study how a concept that is extensively covered in the literature, namely the outbreak of armed conflict, applies to nomadic groups. In turn, there is value in knowing what determines the outbreak of violence, as other future nomadic conflicts may be prevented with more insight.

Literature review

The existing literature does provide a multitude of possible theoretical explanations for the emergence of armed conflict with regard to ethnic groups in

general

. Although the research question posed earlier does not appear to have been

formulated in the present scholarship, the following review should be helpful in several ways. It allows for the research question to be placed in an appropriate context of existing research, aside from offering an initial body of knowledge that can be considered the starting point on which to elaborate. The studies presented subsequently are borrowed from a range of both general, more theoretically-inclined research on the onset of conflict, as well as case-specific knowledge with regards to the Tuareg.

(13)

Theoretical literature

In the 1959 publication ‘Man, the State and War: A Theoretical Analysis’, ’Kenneth Waltz describes three levels of analysis - or ‘images’ - in relation to the question of why a war occurs. From most to least abstract their focuses are; human nature, the internal characteristics of state, and structural and systemic pressures (Waltz, 1959). As mentioned previously, the emphasis of this paper will be placed on the latter. Waltz continues to develop the idea of the anarchical world as a context for war, by which he means the inexistence of sovereign supranational governing bodies (_Ibid.

). Although interesting, its level of abstraction leads to a reduced relevance from

here on.

In the book ‘Understanding International Relations’, Brown and AinIey write that a disruption in the balance of power is a primary cause of war (Brown & Ainley, 2005). The idea of balance of power can be vague, yet at times helpful in understanding processes. _{However, in a more practical sense, they comment on civil} wars and internal conflict that “in principle, states have methods of [internal] conflict resolution which forbid the use of force; sometimes a problem emerges which cannot be contained by these mechanisms, and violence – civil war, if on a large enough scale – ensues as a result” (_Ibid.

, p. 105). Finally, they reiterate Karl von Clausewitz’s

famous line that war is ‘a true political instrument, a continuation of political activity by other means’, thus implying that there is a rational purpose behind systematic violence (_Ibid.

).

James Fearon and David Laitin link the onset of violence to the degree of poverty experienced by a country in their 2003 publication titled ‘Ethnicity, Insurgency, and Civil War’. They use gross domestic product (GDP) per capita as an indicator of wealth, and consequently, as a proxy for measuring likelihood of the outbreak of violence, which although theoretically applying primarily to anti-colonial wars, should have significant relevance when it concerns a wider variety of types of wars (Fearon & Laitin, 2003). Although the indicator does not directly account for the causes of the outbreak of armed conflict, it does point towards some significant underlying influence.

(14)

The first of several leading, more practical perspectives on the eruption of ethnic violence is the 2010 publication by Cederman, Wimmer and Min, titled ‘Why do ethnic groups rebel?’. The paper argues that there are three significant factors that are to be considered when studying the underlying causes of rebellion instigated by an ethnic group (_Ibid.

). The first of these is ‘exclusion from state power’, which is

understood as a lack of representatives or other form of influence in government (Cederman _{et al.}

, 2010). Especially recent losses of power - the worsening of this

factor - is seen as an influential cause of conflict, as it signifies the few alternative means for a group to ensure that their needs are met (_Ibid.

). Second is the

‘mobilisation capacity’, by which the authors intend to measure the extent to which an ethnic group is able to muster, arm, and provide for forces to fight for their cause (_Ibid.

). Lacking this capacity generally ascertains that any efforts are indecisive.

Finally, the authors consider whether an ethnic group has previous experiences of conflict, most notably that which turned violent. The familiarity with violence amongst a population is thought to have the effect of lowering the threshold to which circumstances must deteriorate for violence to occur in future conflicts (_Ibid.

).

Another perspective that should not remain unmentioned is that of Collier and Hoeffler (2004), as described in their paper ‘Greed and grievance in civil war’. In this theoretical discussion on the underlying causes of civil wars, two general categories of reasons for people to mobilise and join an armed conflict are outlined and explored. As the title of the paper gives away, the first of these categories is ‘greed’, which is understood as “atypical opportunities in building a rebel organisation” (Collier & Hoeffler, 2004, p. 563). Greed is a pull-factor, meaning that joining the conflict offers enticing possibilities. The other type of motivation in relation to mobilisation that the paper features consists of ‘grievances’. Grievances are push-factors, stemming from unbearable civilian circumstances, _{pushing people to} mobilise and into the role of soldier. Collier and Hoeffler provide three examples, namely; “high inequality, a lack of political rights, [and] ethnic and religions divisions in society” (_Ibid.

). Although this paper emphasises individual decision-making, it does

present insights which are transferable to the group level of analysis, as it concerns the approach of this paper, and can therefore be referred to in later stadia.

(15)

In the second chapter of his book ‘Why men rebel’, Gurr (1970) explains how relative deprivation theory can be used to clarify the causes of a conflict. With the use of the three graphs below, which show decremental deprivation, aspirational deprivation, and progressive deprivation, respectively, he explains how in situations where actual capabilities of a certain population start to deviate from their expectations, the likelihood of violence increases.

Figure 1

: Graphs illustrating from left to right; decremental deprivation, aspirational

deprivation, and progressive deprivation

(Gurr, 1970, p. 47, 51, 53)

The graphs in figure 1 show three examples of when expectations and actual capabilities begin to deviate, namely; where expectations remain constant, but capabilities decrease, where expectations increase, but capabilities remain constant, and where expectations rise continuously, but capabilities fail to match its rate of increase. Thus, it may be a helpful, though at times oversimplified, analytical tool when these two values can be determined.

Empirical research of non-Tuareg cases

In his book ‘Nomads and the outside world’, published in 1994, Khazanov discusses the idea of nomadic states, saying that when “a nomadic state emerges only [to have] specific relations with the outside world [...], it cannot exist for any prolonged period just because of the internal development of the nomadic societies”, meaning that he sees nomadic groups as intrinsically incompatible with sedentary societies (Khazanov, 1994, p. 296). The alternative he offers for nomadic states to become more sustainable in the long term is for them to “incorporate within itself a part of this outside world in the form of its sedentary population or, either directly or indirectly, it must subjugate that sedentary population” (_Ibid.

(16)

alternative “all such states are not completely nomadic states in the strict sense” anyhow (_Ibid.

). To sum up what this means, sustainable rule be only be achieved by

purely nomadic states when relying on the subjugation of sedentary populations, which may possibly cause conflict.

Evidently, there is no general theory on nomadic violence eruption. Nevertheless, a brief look at some other cases of pastoral nomadic violence could offer some insight.

First, in Tibet, “policies resulting in decreased mobility may exacerbate the negative effects of climate change because of a positive feedback between climate and negative density dependence” is said to create dire circumstances for nomadic tribes, raising the likelihood of violence to occur (Warg Næss, 2013, p. 123). Other sources suggest that “probably the most frequent incitement to war was loss of community livestock to raiders, especially if the action were accompanied by attack and the wounding or killing of anyone” (Ekvall, 1964, p. 1124). Besides, “trespass, or violation in a number of forms, was the next most common cause of war between communities” (_Ibid.

, p. 1124-1125). Trespass is understood as an infringement of the

right to certain pastures. In practice, “trespass and disregard of grazing rights might range from the more or less accidental use of grazing areas under dispute along vaguely marked borders, to flagrant and large scale [devastation] of carefully preserved stands of winter hay near the winter quarters of the community” (Ekvall, 1964, p. 1125). Finally, and in similar fashion, armed violence may take place following the violation of a tribe’s ‘water rights’, which are meant to ensure access to fresh drinking water for the herds (_Ibid.

).

A paper by Ofuoku (2010) on violent occurrences between sedentary farmers and nomadic cattle herders in neighbouring Nigeria (vis-á-vis the Tuareg) titled; ‘Causes, effects and resolution of farmers-nomadic cattle herders conflict in Delta State, Nigeria’, suggests that conflict can arise from the obstruction and destruction of each others means of productions. It is tested that “the causes of such conict were destruction of crops, contamination of streams by cattle, zero grazing of land, disregard for local traditional authorities, female harassment, harassment of nomads by host communities youths, indiscriminate bush burning, defecation of cattle on roads, cattle theft, and straying of cattle” (Ofuoku, 2010, p. 33). Indeed, the paper

(17)

shows, these “socio-economic effects of the conicts result [in] reduction in crop yield and income of farmers/nomads displacement of farmers, loss of lives and properties and loss of products in the storage” (_Ibid.

).

Finally, a report by Fratkin titled “Ethiopia's Pastoralist Policies: Development, Displacement and Resettlement” mentions that there are “cases of violent conflicts between government forces and armed members of the Bodi and Mursi communities, who vow to resist rather than give up their way of life and food production strategy” of nomadic pastoralism (Fratkin, 2014, p. 107).

What seems to be an underlying theme in these other cases, is that the nomadic pastoral lifestyle and means of production is under threat, which subsequently contributes to sparking violence aimed at resisting the consequences of such restrictive circumstances.

In relation to these stresses, another example by Crawford and Leonard describes how; “today, population growth, environmental degradation and changing pattern of land-use threaten many pastoral populations, in part, because it is a subsistence regime that requires large areas of land to support relatively small populations” (Crawford & Leonard, 2002, p. 4). Additionally, “the creation of new political borders, disrupting the traditional migratory routes can have dire consequences for the survival of the nomadic pastoral way of life” (_Ibid.

).

The authors also raise the point that in East Africa, _{“‘traditional’ pastoralist} strategies [have lead] to irreversible environmental degradation, not only in pastoralist zones themselves, but also in more economically valuable regions abutting them”, which indicates that restrictions can in some sense be unintentionally self-imposed _(Ibid.

, p. 100)_{. Among the most catastrophic of these factors is}

“unrestricted growth of livestock herds, which increases the risk of overgrazing, and migratory herding, which expands its scope” (_Ibid.

)_.

The Tuareg case

As may be expected, the Tuareg case has been the subject of earlier investigations, although often times those do not appear to place them in the wider context of nomadic tribes, with the exception of a few. Studies have often provided

(18)

ethnographic research, generally considering the violent episodes experienced by the Tuareg as one chapter of a longer history and placed within a larger culture. These different pieces of literature will be laid out and consulted in detail in the data collection phase.

Theory

While the literature review introduces a series of helpful perspectives, ideas, associations, and historical background, there is a lot left to be desired when it comes to the specific case of the emergence of postcolonial Tuareg violence. Most, if not all, of the papers concerned with civil war onset do not have the intention to delve into the effects of the altered circumstances that arise when dealing with nomadic populations. On the other hand, the research that is devoted to nomadic populations, or more specifically the Tuareg, generally struggle to proceed past a rather superficial discussion of the emergence of armed conflict, perhaps considering that endeavour too niche. Thus, as I transition from literature review towards formulating the hypotheses, I conclude that there is a gap in the literature regarding the sedentary-nomadic divide, which will constitute the core interest of the paper.

Whilst the literature by Cederman _{et al.}

, Collier and Hoeffler, and Gurr in

particular offer a solid starting point for analysing the causes of conflict, this study shall also set out to identify and investigate additional factors contributing to the start of violence with regards to nomadic people. Hence, it will be more concerned with the clues - or hunches - gathered from the various empirical cases of nomadic ethnic groups in conflict. As concluded, a theme shared amongst those populations in Tibet, Nigeria and Ethiopia, is the role of restricting factors to living their pastoralist nomadic lifestyle, in the origin of armed conflict. Through analysing these restrictions, I am to find out whether these are more significant than stated in the literature or if current theoretical approaches are satisfactory.

(19)

The logic of the argument

The logic of the argument that is presented in this paper stems primarily from the empirical findings laid out in the literature review. From these examples the idea is distilled that certain factors limiting the ability of pastoral nomads to lead their traditional lifestyle (x variable) can be a contributing factor to the emergence of armed conflict (y variable). Thus, the argument can be structured as follows; evidence is gathered to show that a) pastoral nomads live a different lifestyle different from sedentary populations, b) circumstances affecting this lifestyle are worsening, and c) this decline is contributing to the start of armed conflict. If found, some degree of causality between x and y can be concluded.

Hypotheses

Deriving from and aiming to add to the literature, the primary objective is to test a general hypothesis that applies to all nomadic peoples. Developed in accordance with the general research question, it goes as follows _{; nomadic}

pastoralist ethnic groups that experience restrictions on nomadisation or pastoralism are more likely to become involved in armed conflict

. The approach to testing this, as

is mentioned earlier, is through a case study of the Tuareg. Consequently, it is deemed helpful - in similar fashion to the RQs - to explicitly state a case study hypothesis. It is expected that; _{the Tuareg became more likely to become involved in} armed conflict during the time since the decolonisation of their home territories due to a combination of restrictions on their nomadic and pastoralist lifestyle.

As is evident from the hypothesis, the investigation will aim to study a range of aspects that are believed to have a possible influence on the origin of violence. Therefore, sub-hypothesis 1 states that; _{the factors influential in causing the}

emergence of armed conflict include, but are not limited to all restrictions on nomadisation and pastoralism that are both man-made - intentional and unintentional, internal and external in relation to the ethnic group - as well as environmentally induced

(20)

The purpose of formulating these factors into a sub-hypothesis is to direct the research to test them as observable implications. The more of these observable implications are found to have a significant impact on eruption of violence, the stronger the evidence for the main hypothesis is deemed to be. Therefore, some examples of factors that shall receive attention include the imposition of national borders, anti-nomadic government policies, as well as climate change and desertification.

Considering that the above hypotheses mainly relate to the causal effects, a second sub-hypothesis is devised to address the expected causal mechanism. As such, sub-hypothesis 2 predicts that; _{a combination of one or more of the factors} described in sub-hypothesis 1 can accumulate towards a threshold, at which point the nomadic group´s perceived best alternative approach to sustain their lifestyle is to instigate armed conflict, in order to increase its agency.

At this point, it seems necessary to define want is meant by both ‘armed conflict’ and ‘violent conflict’, which are sometimes taken to have slightly varying meanings. These terms could carry connotations of scale and types of parties involved, and so, the difference between violent and armed conflict requires addressing. Violent conflict is defined by Wallensteen and Axell as “the use of armed force by two parties, of which at least one is the government of a state, that results in at least 25 battle-related deaths” per year (1994, p. 333). Armed conflict then, is defined as a “conflict between two or more parties, but usually between the state and an anti-state group or groups using weapons on a sustained basis” (Payson Conflict Study Group, 2001, p. 8). Comparing these definitions shows that _{armed conflict is} the broader of the two, not limiting the number of parties and being less demanding of scale minimums. The intention is to not limit the investigation to conflict with merely two contesting parties, yet the minimum casualty count appears necessary to include, ensuring that discussed events actually are deadly conflict, which is crucial to the subject of the study. Thus, the 25 casualty threshold shall be considered as much as possible, whenever reliable data is available. Having discussed this, from here on the two terms are used interchangeably.

(21)

Alternative theories

Although the hypotheses presume a particular explanation to be more likely than others, it is always possible that another theory is better able to account for an observed phenomenon.

One alternative theory of causes of nomadic conflict has been mentioned earlier, namely, that it is in their nature to be more violent that sedentary populations. “Nomads are described as ‘natural soldiers’, since their pastoral way of life enabled them to acquire individual military skills (such as riding, shooting, endurance) from an early age, but the effectiveness of this skilled army varied greatly over the years” (Biran, 2004, p. 341). This idea however, is deemed to be rather superficial, and so, seems less likely to reliably explain, for instance the Tuareg rebellions. If, during the data collection, convincing information surfaces, signified by observable implications such as unprovoked violence and violence amongst all nomadic peoples, then further testing of this hypothesis could be considered.

Secondly, the variable degree of commitment to nationalism is an interesting one to consider amongst nomads. Intuitively, pastoral nomads would not be expected to hold particularly strong pro-state sentiments. While some authors (Brown, 1997) suggest the influence of nationalism during the onset of armed conflict, it assumes stationary ethnic groups, not transboundary nomadic groups, which may experience different nationalist situations in various nations.

Arguably, the reason why existing theories (e.g. Cederman_{et al.}

, 2010; Collier

& Hoeffler, 2004; Gurr, 1970) do not differentiate between nomadic and sedentary populations is because no significant distinctions between the two have appeared to those authors, when it comes to the emergence of conflict. In such case, the present literature would not require any elaboration on the subject other than the notion that indeed, the residential status of a group is of no particular interest to the onset of violence. However, it should be noted that these theoretical perspectives contemplate the causes of conflict on a more abstract level, and thus consider the average causes and effects. Thus, the nomad aspect could be an outlier to their theories. These questions will be reevaluated post testing.

(22)

An obvious obstacle to the hypothesis stems from the question; did pastoral nomads not get into conflict before the lands they inhabited were colonised, when they lost self-determination and borders were drawn that restricted their mobility? Indeed, this does seem quite plausible. The greed factor, as outlined by Collier and Hoeffler (2004), could offer an answer to the relative lack of grievances during earlier periods. Historically Tuareg raids of neighbouring populations have been an influential cause of violence. In fact, “the economic basis of ‘nomadic’ social and political units normally encompasse[s] a variety of types of production, of which pastoral production was the most important but not the only one” (Di Cosmo, 2002, p. 168). Di Cosmo mentions that because of most nomadic peoples’ incomplete self-sustainability, either trade with sedentary populations or raiding them has often been a necessity (_Ibid.

). On the other hand however, “raids […] were small-scale

enterprises until strong sedentary states imposed a new order on the frontier, making it more secure” (_Ibid.

, p. 169). In response, Di Cosmo continues to explain, nomads

would form their own unions and corresponding supra-tribal politics, able to rival these new states in order “to pressure [them] into yielding to their economic demands” or extortions (_Ibid.

). Arguably, such developments do still fit the hypothesis

- expecting restrictions on nomadisation to drive nomads to conflict -, provided that one understands not only pastoralist production to be intrinsic to nomadisation, but raiding and trading as well. With that in mind, this organised resistance to raids can be regarded as obstructing nomadisation. All in all then, depending on whether someone considers this to be an example of greed of improving nomadic circumstances, or overcoming resistance to nomads practicing their traditional lifestyle, the theory could be adjusted to incorporated. Regardless, the focus of the investigation will remain on restrictions, not opportunities.

Counterevidence

It is imaginable that restrictions on nomadisation do not lead to conflict. Examples of this may include, firstly, desertification weakening and starving nomads, through the proxy of their herds, to the degree that violent opposition to any other party ceases to be a priority, and secondly, harsh restrictions on their lifestyle could

(23)

motivate pastoral nomads to reconsider their alternatives and simply opt to settle in one location. These ideas could illustrate how restrictions may not always result in violence.

If the hypotheses are not correct or another explanation is better able to explain the violence among the Tuareg, I expect to find sources explicitly citing different causes and disregarding those restrictions as significant. Moreover, I predict that there is no consistency in the time between restrictions on pastoralism and nomadisation and conflict (i.e. random lag between the independent and dependent variables).

Another scenario exists where the hypothesised causal factors are observed to be present, as well as their proposed effects - i.e. a correlation between x and y -, despite the two not being causally related. The observations would be the same, yet, one would expect to find that in exploring the causal mechanism the two variables are not linked to each other.

Finally, if the exact opposite of the hypothesis is correct, then fewer restrictions on nomadisation should lead to increased probability of conflict. The first of several possible explanations would be that weak state governments, which are less likely to impose restrictions, are less able to resist raiding by nomadic tribes and less able to suppress grievances, both spurring violence. In other words, both greed and grievance factors may arguably allow for increased violence in the context of fewer restrictions.

If any of the above is concluded, it would imply that there is no significant difference in the developments that lead to violence in nomadic tribes as compared to sedentary populations. This would restate the strength and relevance of existing literature.

Research design

First of all, the data collection is based entirely secondary research, making use of a variety of published literature including peer reviewed papers, books on Tuareg history and of Sahelian countries, news articles and other sources deemed

(24)

credible. The information gathered from this combination of sources is synthesised into a systematic response to the research questions.

Case study selection

The design of the research is based, first and foremost, on the use of a case study. In order to test the hypotheses, the case that is being observed in greater detail is that of the spells of armed Tuareg revolt. Although their lands are sparsely populated, these people cover an area three times the size of modern-day France (Keenan, 2003a), located primarily in Mali and Niger, although they are present with significant numbers in Libya, Algeria and Burkina Faso (de Laszlo, 2015). Nevertheless, the focus of this paper lies on the former two countries, acknowledging that the vast majority of the conflict appears to have taken place within their current national boundaries. This decision means that I select a case based on dependent variable, presence of armed conflict. The Tuareg case, as it will be referred to, is restricted to the time period since the decolonisation of west Africa and split from the French Empire. The dates of independence differ amongst these countries, but lie approximately around the start of the 1960s, plus or minus several years.

In selecting the case, several considerations were made in regard to the timeframe in which the conflict ideally took place. With the assumption in mind that more recent secondary literature, especially that concerning a relatively remote ethnic group as the Tuareg, is likely to be more accurate, historic cases - predating the 20th century - were ruled out. However, conflicts that emerged recently are oftentimes still shrouded in confusion and lack recognised publications like peer-reviewed articles on the matter. Although the Tuareg violence occurs with regularity to this day, a useful body of knowledge exists on the period since around 1960. Regarding the ongoing conflict, it is decided to include the episodes of organised violence up until 2013, which, as will be argued later on, is when the most recent rebellion ended.

(25)

Variation in the dataset

Since the start of the 1960s and the collapse of French colonial West Africa, multiple episodes of Tuareg rebellion have taken place. Given that each rebellion can be tested for its underlying causes, the dataset consists of events that occurred over a range of approximately half a century. This temporal variation is an unavoidable element that arises from the decision to compare a single ethnic group in different periods in time. Although this could lead to some additional inaccuracy, it may also offer insight into the relevance of the hypothesis at different stages in recent history. Despite an initial concern being an excessive amount of time to be studied, the intermissions of peace account for much of those 50 years, ensuring that the dataset remains manageable.

Moreover, there is considerable degree of spatial variation in the Tuareg case study. Although dealing with a contiguous geographical area, the events (the outbreaks of violence) occur most notably in Mali and Niger. The large domain of the Tuareg brings to the attention another type of variation held within the dataset. There is a noteworthy degree of variation between the subgroups within the ethnic group, which may influence which restrictions they experience and how. Note that this is no reason to consider them as different ethnic groups. Similarities as well as differences will be explored in further detail in the ‘results section’.

Advantages and disadvantages of the case

With the aim to explore both the causal effects and mechanisms as they relate to nomadic communities in a wider sense, the Tuareg case is perceived to be an appropriate subject of analysis, as it should allow for the inquiry into both categories of causality. Is should be mentioned however, that the selection of a single case study limits the confidence with which conclusions can be used to make inductive inferences about all nomadic peoples. In other words, causal effects are described and indexed rather than generalised.

The Tuareg case was chosen above other possible options for several reasons. In a practical sense, it was believed that the contemporary and ongoing

(26)

nature of this conflict, without reducing temporal variation (i.e. it has lasted relatively long), would be conductive to the collection of data, more so than conflicts that have been resolved many decades ago.

A decisive factor in electing this particular ethnic group, is the documented existence of both the independent and dependent variable; the restrictions on nomadisation and armed conflict, respectively. What this signifies, is that the chance of the case study providing relevant insight into the causal mechanism, is significantly higher than when choosing a nomadic ethnic group at random, or on the basis of solely the dependent or independent variable.

Dependent variable

The dependent variable - or Y variable - is measured as: an increase in the likelihood of the outbreak of armed conflict involving nomadic ethnic groups. Again, the definition of armed conflict here will incorporate elements from the conflict definitions reviewed earlier, namely that at least 25 battle-related deaths occur annually (Wallensteen & Axel, 1994) and that it may concern two or more parties, governmental or not (Payson Conflict Study Group, 2001). The 25 metric should suffice, given the low population density in these parts of the world, to label violence as an armed conflict. In order to study dependent variable, descriptive inferences are made in the first section of the results, regarding the years in which armed conflict took place and, most importantly, began.

Then, considering conflicts exclusively between non-governmental parties, in addition to those where a government is involved, could lead to additional observable implications, namely whether violence is directed at the party most closely related to the origin of a restrictive factors.

Independent variable

The independent variable - or X variable - is measured as; the emergence of (a combination of) restrictions on nomadic and pastoralist lifestyles over time. As hypothesised, tt is expected that there is considerable variation in the types of restrictions. Examples include, but are not limited to sedentarisation government

(27)

policy, national borders cutting of territory, territory seized by others (e.g. for resource extraction), overpopulation, overgrazing and drought. Data is collected on the basis of qualitative analysis of each period that predates the onset of violence. This is considered a valid method, as it omits abstracted indicators, and reviews a process supported by in-depth analysis of the causal mechanism. The data is broadly divided in two categories: restrictions on nomadisation and restrictions on pastoralism.

The most prominent limitation lies in the difficulty of adding up the severity of multiple qualitative assessments of restrictive factors. This will have to happen on the basis of informed estimates and judgements of each period.

Finally, it is conceivable that some factors are restrictive in the short term, but facilitate or promote nomadisation in the long term. However, this ambiguity is not deemed to be overly obstructive to the investigation, given that the combined severity is considered, allowing for fluctuations in individual restrictive factors. Moreover, if violence breaks out before such a factor changes to become favourable to nomadism, it would not contradict the hypothesis.

Control variables

The rationale behind choosing several different rebellions initiated by the same ethnicity against the same government or sedentary populations, is that it controls for many different variables, which include; cultural and ethnic differences amongst nomads, unreasonably wide spatial variation, and unreasonable differences in governments and sedentary populations. Violent periods in Burkina Faso, Libya and Algeria, limited as they are, would introduce three additional governments. It is believed that included these countries would require a breadth of research that exceeds the scope of this investigation. Finally, as with any non-experimental study, and especially when a cross-border conflict (between Mali and Niger), there will always be variables that cannot be overseen, let alone be controlled. Their influence shall be evaluated and require some reflection in the cautiousness of the conclusions.

(28)

Observable implications

In the hypotheses, I predict that there are several sources from which restrictions on nomadisation can originate. Based on various origins of the restrictions that are thought to contribute to the outbreak of violent conflict, I expect variation in the manifestations of the violence. What follows is a brief overview of possible observable implications, which may be used to deductively make inferences with regards to the nature of the causal effects as well as mechanisms.

The pattern expected to be observed, first of all, consists of spatial and temporal variation in the outbreaks of violence in a positive correlation with time and place where the restrictions on nomadisation have experienced an increase in severity.

Similarly, it is expected that those subgroups of the Tuareg population that suffer the harshest restrictions are more likely to exhibit violence. This translates to a situation where the independent variable is expected to be measured to be the greatest. In practice, this means that the circumstances of the subgroups responsible for initiating the violence should be tested with particular interest.

Depending on whether or not the restrictions are human-induced, violence is predicted to be directed more intentionally towards the actor or actors responsible. The reasoning behind this, is that a made-made restriction, especially when introduced intentionally, may inject an element of retaliatory sentiment into the situation, targeting whoever is regarded as the source of the increased hardships. To illustrate, consider how intentionally government-imposed restrictive policy would seem more likely to spur violence towards this government’s forces, than it would towards a neutral third party.

In as far as it is possible with a conflict that seems to be ongoing and escalating with regularity, it would seem interesting to consider the points that are addressed in peace talks. Although the investigation emphasises the starting phases of violence, it is likely that there are clues hidden about the onset in negotiations leading to the resolution of the conflict. The logic behind this lies in the idea that

(29)

representatives of the Tuareg will be intent on framing the underlying reasons for them to fight are framed as their priority in such talks.

Types of variation if then

Spatial Restriction manifests in location _A

Population of location _A are more likely to become violent

Temporal Restriction begins to manifest during year _t

Violence in expected year

t or later (within a reasonable period) Degree of restriction experienced Restrictions target population _α more than average Population _α is more likely to become violent

Origin of restriction Restrictions are caused by actor(s) _β

Actor(s) _β

are more likely

to suffer violence Negotiation demands Tuareg negotiators

demand resolution to restriction _x

Restriction _x

is more likely

to have contributed to the violence

Table 1

: Summary of the observable implications

The case study template

In studying the case, the objective is to identify events where violence started. Each of these events will be approaches using a set of guiding questions that have the objective of deriving comparable information from those various instances.

- When did the violence occur? - Who committed the violence? - Who was targeted by the violence? - Where did the violence take place?

(30)

- What was the aim of the violence?

- Which restrictions on nomadisation or pastoralism may have caused violence? - Could there be other reasons for violence besides restrictions?

- Is there anything notably deviating or special about the occurrence? - What was the result?

The rationale behind systematically posing a predetermined set of questions, is to be able to understand how this secondary research was conducted, to be able to reproduce it, and to be able to apply it to different cases of nomadic conflict.

Level of analysis

In approaching the investigation, the level of analysis lies on the scale of the group more than the individual. To illustrate what this entails, the explanation for individual people to enter into a conflict are likely to emphasise different aspects of the causal mechanism as compared to group mobilisation. The former may speak of personal reasons and circumstances, whereas the latter aims for a certain degree of generalisability among a group. Nevertheless, this should not imply that individuals are entirely disregarded. People in leadership roles may prove to be of interest for gaining insight in the group dynamic. For instance, through ongoing “political organisation along ethnic lines [or] ‘ethnic mobilization’; leaders decide to speak for ‘their’ ethnic group, thereby making the abstract idea of ethnic belonging a somewhat more tangible reality, and engage the members of this group into political action” (Vermeersch, 2011, p. 1). As such, some individual-level processes are relevant for the group level as well. Nevertheless, the research question is framed to prioritise the Tuareg as a population, not each member individually.

Ethical considerations

Given the nature of secondary research as conducted in this study, exclusively publicly available information will be analysed and used to draw conclusions. Thus, no person that identifies as part of the case study group or other is consulted directly. Consequently, no unpublished sensitive information about any person or group is

(31)

added to the world’s collective knowledge. In sum, there are no significant ethical concerns raised by the specifics of the research design.

Results

Section 1 - Background and descriptive inferences

Introduction

The Tuareg, or Kel Tamasheq in their native language (literally; people who speak Tamasheq), is a North African berber ethnic group of pastoral nomads, which is speculated to number anywhere between one million and 2.5 million, the latter being based on aggregated data from the CIA World Factbook (de Laszlo, 2015; Pike, 2015; CIA, 2006; CIA, 2010; CIA, 2013 (although other sources suggest ranges as wide as between 100,000 and 3.5 million)). The Tuareg, as they shall be referred to from here on (with the widespread use of French, Tamasheq, as well as English names of people, places, organisations, etc., the English will be used by default, unless otherwise specified), have been living in the central/western Saharan regions for many centuries.

Western romantistion of the Tuareg has often been linked with the “indigo dye from their turbans, an age-old mark of the Tuareg that led early visitors to dub them the ‘blue men’” (Gwin, 2011). “By tradition Tuareg men cover their faces, though the women do not. The layers of cloth not only protect from the harsh sun and wind but also conceal their emotions”, which has added to the mysticism that surrounds this ethnic group in the minds of many who are little familiar with its background (_Ibid.

).

Simultaneously, this has contributed to the fame and partial representation of these people in other parts of the world. As such, it is helpful to provide some background.

Tuareg history

Some of the first records of the Tuareg as an ethnic group suggest that they have lived in the central Saharan region since the 5th century, including that of the

(32)

Greek historian Herodotus, albeit referring to an older name; the Garamantes (Douglas-Bowers, 2013). According to folklore, this period began during the life of their ‘Founding Queen’, Tin Hinan. During the 11th century, the Tuareg founded the famous city of Timbuktu and a period of consolidation of power in the region followed, in which many of the caravan routes came into existence (_Ibid.

). The Tuareg

converted to Islam in the 14th century, primarily for more favourable relations with islamic trading partners, allowing them to attain considerable wealth (_Ibid.

). This

changed under the colonial rule that followed.

The French colonisation of West Africa did not immediately encompass the Tuareg, with extremity of the environment playing a large role in their consideration (Schoenahl & Whittington, 2003). Only in the late 19th century, when competition from the British and Ottomans became apparent, did the French decide to expand into the interiors of West Africa (_Ibid.

; Idrissa, 2003). Significant resistance was posed

to the French invasion and led to several massacres on both sides. After the Tuareg were finally beaten with help of the advancements of French weaponry, treaties were signed in 1905 in Mali and 1917 in Niger (Schoenahl & Whittington, 2003).

Until the last resistance was quelled the colonial administration acted through highly centralised government and enforced constraining policies with regards to their livelihoods. “By requisitioning large numbers of Tuareg camels, the French were able to administer the Tuareg lands they occupied, while also undermining the ability of the nomads to mount an effective resistance” (Pike, 2015). These requisitions, in combination with camel losses due to the drought that lasted from 1911 until 1914 “undermined the basis of the desert economy -- animal husbandry, the salt and animal trades, and income from providing transport services” (_Ibid.

). In some places

the camels lost was more than half the total number: “at the time of the 1912 annual caravan, 16,821 camels were recorded in Bilma and 6,721 in Fachi as compared to 8,220 and 2,410 respectively in 1914” (Idrissa, 2003, p. 197). This is said to be the main cause for rebellions to emerge in various Tuareg regions, including most notably the Koacen revolt in the Aïr mountains, but also in Azawad, and by the Niger river bend (Pike, 2015). The resistance lasted roughly until 1917 or 1918, but the effects of the repressive governance approach and drought crisis continued to be felt by the Tuareg population.

(33)

It should be noted that in this investigation this period shall not be subjected to further examination. Even though the revolt against the French is regularly mentioned when discussing the subsequent, post-colonial violent spells, it is believed that testing the violence of the early 20th century would unnecessarily introduce additional inaccuracy. Since, all of West Africa was under colonialism during this period, making comparison would be troublesome as there is a lack of control variables. Nevertheless, it seems that restrictions on nomadisation and pastoralism indeed played a role in these conflicts, and thus adds to the legitimacy of the hypothesis.

Independence came for West Africa around the start of the 1960s, with some variation between the newly self-ruling countries that emerged. The fall of the empire did create a certain power vacuum. Posen (1993) conceptualised this situation of the collapse of an empire as a security dilemma, in which ethnic groups that lived in it are “compelled to provide [their] own protection [and] must ask the following question about any neighbouring group: is it a threat? How much of a threat? Will the threat grow or diminish over time? Is there anything that must be done immediately? The answers to these questions strongly influence the chances for war” (Posen, 1993, p. 27). As such, the collapse of the French colonial empire foreshadowed the rebellions to come.

Indeed, the Tuareg “_{were motivated to imagine a post-colonial dispensation.} Many Tuaregs in Mali (and neighboring countries) had begun to dream of an independent state - Azawad - comprised of Tuareg-populated territory in northern Mali, northern Niger, and southern Algeria_{” (Pike, 2012a). Nevertheless, priority} always lay with the local community, and unity among the ethnic group as a whole was not demonstrated, neither politically nor militarily (_Ibid.

). At this point in time, I

conclude the historical overview and transition into the rebellion phases.

Differences and similarities

The ethnic ties between members of the Tuareg population are seldom said to be derived from shared genetic origins. These mixed genetics hint at “one of the riddles of the Tuareg, who have always considered themselves a people apart yet for centuries took slaves from other desert tribes and intermarried with them” (Gwin,

(34)

2011). Perhaps this ambiguity also added to the mystery. Although skin colour has historically been exploited to stratify Tuareg society into a class system, both through internal processes, but also through colonial rule, it has not caused insurmountable divides (Lecocq, 2010). Either way, “the result is an ethnic group distinguished primarily by its common language”, Tamasheq, which is cited as the main factor from which Tuareg identity is derived (Gwin, 2011). Consequently, maintenance of the Tamasheq language is an important shared cause.

Despite their similarities, the Tuareg are simultaneously divided on several subjects as well. One of those is religious dedication, which has some significance to the last phase of rebellion that I will discuss. However, more importantly, the Tuareg as a group are undecided with respect to their stance on nomadisation and pastoralism itself.

“_{Most Tuareg want to see this natural beauty, this freedom of the} wide open spaces preserved and with it the nomadic pastoralism that has been practised there for millennia. Then again, there are some, a few, that consider nomadism to have no future at all and who urge their fellow Tuareg to accept the sedentary life as the only route to a modern and sustainable future

” (Morgan, 2014).

Consequently, the latter group is appears to be more in line with the stance of the governments of Mali and Niger, desiring their assimilation. Therefore, it is expected that this divide is related to the degree in which these varying Tuareg are willing to rebel and fight for independence.

Nevertheless, the above is not in contradiction with the notion that practically all Tuareg are united in their grief regarding the impediment of the social and economic development of their homeland since independence from France (_Ibid.

).

“They would like to see more schools, more health clinics, more wells, better roads, cheaper petrol, better food distribution, less criminality, more peace and stability. [...] Without development, the desert is going nowhere” (_Ibid.

). This idea is widely shared,

(35)

Spatial variation - Tuareg confederations

Historically, there have been seven Tuareg confederacies, or ‘drum groups’, which include the Kel Adrar, Kel Atazam (Ioullimiden, west), Kel Dinniq (Ioullimiden, east), Kel Gress, Kel Aïr, Kel Ahaggar and Kel Ajjer (bote that many different spellings of each of these names are used, depending on various transliterations). Figure 2 shows their approximate distribution within the combined Tuareg domain. Each of these independent confederacies is led by an elected chief, or ‘Amenokal’ (literally; owner of the land), in coordination with a counsel of different smaller federated tribes (Lecocq, 2010). “His precolonial function was to conduct peaceful relations with outsiders or to lead expeditions against enemies; [since colonisation] he acts as a liaison with the central government” of the French and later Malian and Nigerien administrations (Levinson, 1995, p. 368).

This type of political organisation is mostly the result of the “vast desert habitat and dispersed nomadic lifestyle, both of which tend to place allegiance to blood and tribe above allegiance to nation or ideology and militate against collective thought or action” (Morgan, 2014). In other words, an individual Tuareg person is likely to prioritise his confederacy over the Tuareg ‘nation’. This has sporadically resulted in disharmony and even conflict between the confederacies. Below in table 2, the various historic divisions are summarised.

(36)

Figure 2:

French language map of showing the distribution of the various Tuareg

confederacies (Kel Adrar, Ioullimiden/Kel Atazam, Ioullimiden/Kel Dinniq, Kel Gress, Kel Aïr, Kel Ahaggar and Kel Ajjer

(Rekacewicz, 1995) (note that the Kel Tademekkat

pictured on the map in Mali is rarely mentioned in the literature and dissolved prior to French conquest (Lecocq, 2010), and is therefore not included)

Confederacy number Name Location

1 Kel Adrar (or Kel Adagh,

Ifoghas)

Kidal, Timbuktu and Ifoghas mountains (Mali)

2 Ioullimiden/Kel Atazam Menaka and Azawad

region (Mali)

3 Ioullimiden/Kel

Dinniq

Tchin Tabaradene, Abalagh, Teliya and Azawad region (Niger)

4 Kel Gress Zinder and Tanout (Niger

and northern Nigeria)

From herd to horde : the emergence of nomadic pastoralist armed conflict : a Tuareg case study

Acknowledgements

Abstract

Contents

Figures and tables

Introduction

Definitions

Research questions

Approaching the research question

Literature review

Theoretical literature

Empirical research of non-Tuareg cases

The Tuareg case

Theory

The logic of the argument

Hypotheses

Alternative theories

Counterevidence

Research design

Case study selection

Variation in the dataset

Advantages and disadvantages of the case

Dependent variable

Independent variable

Control variables

Observable implications

The case study template

Level of analysis

Ethical considerations

Results

Section 1 - Background and descriptive inferences

Introduction

Tuareg history

Differences and similarities

Spatial variation - Tuareg confederations