All of us know of the terrible violence unleashed in Darfur and on Darfur in 2003-04.

That violence has a history. It also has a meaning. Violence is not its own explanation. When people fight, whether non-violently or violently, there is inevitably a history and inevitably issues around which they fight.

The violence in Darfur began as a civil war in 1987-89. There was a reconciliation conference at the end of that civil war. Both sides made representations at this reconciliation conference, putting forth their point of view on the conflict. Both claimed to be victims.

Human-rights organisations from Human Rights Watch to the International Crisis Group have focused on the atrocities committed during the violence in Darfur.

I want to focus on the background that they have ignored: Why the violence? What were the issues that drove the civil war? The violence in Darfur is usually described as an ethnic conflict, sometimes even as a racial conflict.

The question is: Do Africans fight one another just because they are different? Or do they fight because they have differences?

The background to the conflict in Darfur is marked by two most important issues.

The first, the more immediate though not necessarily the most important, is that of sheer survival in the face of an ecological crisis, a crisis of drought and desertification.

According to a UNEP study issued last year, the Sahara has moved roughly 100 kilometers in 40 years, pushing northern tribes southwards.

The second issue was even more long-term. It stemmed from the land tenure system created during the colonial period.

Like most places colonised by Britain after the Berlin Conference in late 19th century, Darfur was tribalised by being divided into tribal homelands during the colonial period.

The British divided the tribes of Darfur into three: Settled peasant tribes got the largest homelands, equivalent to their settled areas; semi-settled cattle nomads got smaller homelands that included their villages but not necessarily their grazing grounds; finally, tribes of camel nomads who had no settled villages got no tribal homelands.

To understand the responses of different tribes to the drought that reached its most acute expression in the mid-1980s, you needed to understand this background of how tribal homelands had been created in the colonial period.

When nomadic groups from the north began pushing south, a confrontation built up between peasant and pastoral groups around the lush territory of the Jebel Merra — a mountain in central Darfur.

Each side justified its demands in a different language. The peasant tribes spoke the language of group rights — they defended their exclusive right to the land as their tribal homeland.

In contrast, nomads spoke the language of individual rights: We are citizens of Sudan, citizens have a right to go, to live and to make a living anywhere in the country. Land is not a tribal possession; it is a statutory right conferred – in law – by the state.

To understand the ways in which Darfur resonates with other African contexts, we need to look at the colonial legacy more fully. We shall then be able to understand the challenge we faced at independence, and the extent to which we have proved equal to meeting it.

In most African colonies in the 20th century, the census classified the population into two broad, overall groups.

One group was called races, the other tribes. In some places, authorities distinguished between those said to belong to a race and those said to belong to a tribe; in other places, they distinguished between two kinds of races, between uncivilised races that were further sub-divided into tribes and races that were civilised enough not to be so sub-divided.

It was said that tribe and tribalism would go away with modernity – modernisation, urbanisation, industrialisation, and so on.

Whatever the case, this single distinction – between race and tribe – sums up the technology of colonial governance. I shall explain this with four observations:

First, the distinction between races and tribes was not one between colonisers and colonised, but between non-natives and natives: non-natives were tagged as races, whereas natives were said to belong to tribes.

Races were said to comprise those not indigenous to Africa (Europeans, Asians), or those who were constructed as not indigenous (Arabs in Sudan, Coloureds in South Africa, Tutsi in Rwanda). Tribes were said to be all those defined as indigenous in origin. The state thus distinguished non-indigenous races from indigenous tribes.

Second, this distinction had a direct legal significance.

All races were governed under a single law, civil law. Civil law was full of discriminations — racial discrimination distinguished between master races (European colonisers) and subject races (colonised immigrant groups such as Asians, Arabs, Coloureds and so on).

Subject races were excluded from the exercise of those rights considered the prerogative of members of the master race. But this discrimination was internal to civil law, the body of which applied to all races.

This, however, was not true of tribes and customary law. There was never a single customary law to govern all tribes as natives, as one racialised group. Each tribe was ruled under a separate set of laws, called customary laws.

It was said that tradition was tribal; natives must thus be divided into tribes, with each tribe governed by a law reflecting its own tradition. Yet most would agree that the cultural difference between races — such as Whites, Asians and Arabs — was greater than that between different tribes.

To begin with, different races spoke different languages, mutually unintelligible. Often, they practised different religions.

They also came from different parts of the world, each with its own historical archive. Different tribes, in contrast, were neighbors and usually spoke languages that were mutually intelligible (thus it was said that neighbouring tribes could “hear” one another).

My point is simple: Even if the races were as different culturally as were whites, Asians, and Arabs, they were ruled under a single law, imported European law, called civil law, modified to suit a colonial context.

Even if their languages were similar and mutually intelligible, tribes were governed under separate laws, called customary laws, which were in turn administered by ethnically distinct native authorities.

With races, the cultural difference was not translated into separate legal systems.

Instead, it was contained, even negotiated, within a single legal system, and was enforced by a single administrative authority.

But with tribes, the case was the opposite: Cultural difference was reinforced, exaggerated, and built up into different legal systems and, indeed, separate administrative and political authorities. In a nutshell, different races were meant to have a common future; different tribes were not.

My third observation: The two legal systems were entirely different in orientation.

We can understand the difference by contrasting English common law with colonial customary law. English common law was presumed to change with circumstances. It claimed to recognise different interests and interpretations.

But customary law in the colonies assumed the opposite. It assumed that law must not change with changing circumstances. Rather, any change was considered prima facia evidence of corruption. Both the laws and the enforcing authorities were called “traditional.”

Traditional authority in the colonial era was always defined in the singular. We need to remember that most African colonies did not come from a political history of an absolutist state.

Instead of a single state authority whose writ was considered law in all social domains, the practice was for different authorities to define separate traditions in different domains of social life.

The rule-defining authority thus differed from one social domain to another; besides chiefs, the definers of tradition could include women’s groups in the marketplace, age groups in the battle field, clans when it came to land, religious groups when it came to the spiritual domain, and so on.

The big change with the colonial period was that Western colonial powers exalted a single authority, called the chief, as the traditional authority. Marked by two characteristics, age and gender, the authority of the chief was inevitably patriarchal and authoritarian.

Fourth, this legal project needs to be understood as part of a political project.

The political project was highlighted by the central claim that natives are by nature tribal. In time, this very claim would be advanced as a reason for why the African colonies have no majority, but only tribal minorities.

This claim needs to be understood as political, not because it is not true but because this truth does not reflect an original fact but a fact created politically and enforced legally.

This form of governance had a name in the colonial period. It was called Native Administration. At the heart of Native Administration was an administrative and legal distinction between “native” and “non-native” tribes. Immigrants were identified as “non-natives” no matter how many generations they had lived in the area.

The native identity involved three distinct privileges:

The first was right of access to land. A non-native who wanted access to land in a tribal area had to pay a part of the harvest as tribute to the tribal authority.

The second involved right of participation in the administration of the Native Authority. The leadership of every Native Authority could only be appointed from among those identified as natives.

It was only at the lowest level of administration — the lowest tier of the Native Authority — that “non-natives” could be appointed as headmen in villages where most residents were not natives.

The third privilege was in the area of dispute settlement, for every Native Authority settled disputes using a set of what were called “customary laws.” which inscribed in law the prerogative of natives over non-natives.

The regime of inequality between supposedly original residents and subsequent immigrants led to a tribal administration ruling over a population belonging to multiple tribes.

With all key rights — from access to land, to the right of participation in local governance to rule-making for settling local disputes — defined as group rights and declared the prerogative of those with a native tribal identity, this identity was turned into a fulcrum around which developed, in time, an explosive confrontation between two kinds of residents in every administrative unit — those defined as native and those not.

A situation where an ethnic administration oversaw an ethnic monopoly over land by settling disputes using an ethnically defined customary law was one where administrative power institutionalised ethnic discrimination.

You may have noted that I have shifted between the use of the words “ethnic” and “tribal.” But it is time we distinguished between them.

Whereas ethnic identity is a historically evolved, language-based, cultural identity, tribal identity is better understood as the product of an administratively designated power over a tribal homeland.

Although tribal identity in many cases coincided with ethnic identity, by which I mean language-based, cultural identity, this was not always the case.

In some cases, the same ethnic group was divided into several tribes administratively.

This made it clear that tribe referred not to an ethnic group (language group) but to an administratively designated political community.

Sometimes that political community was created out of a language group; at other times, out of a combination of language groups; at yet other times, it involved a split of one language group into many tribes; just as in some cases “tribe” was designated totally arbitrarily – “invented,” it is said in the literature.

The only thing common between all these cases is that tribe was an administrative unit, and tribal identity an officially designated administrative identity linked to a tribal homeland.

For this reason, I believe it is best to refer to the system of Native Administration and Indirect Rule as a system that institutionalised tribal discrimination, even if it tried to justify this as an inevitable consequence of ethnic difference.

Though I have described this system in the context of Darfur, it obtained in all African contexts that I have studied, from East Africa to Nigeria, and from Sudan to South Africa.

Did tribe exist before colonialism?

Tribe as an ethnic group with a common language did.

But tribe as an administrative entity which distinguishes between natives and non-natives, and defines access to land, participation in local governance and rules for settling disputes according to tribal identity – whether native or not – certainly did not exist before colonialism.

One may indeed ask: Did race exist before racism? As differences in pigmentation, or in phenotype, it certainly did.

But as a fulcrum for group discrimination that signified race difference, it certainly did not. Race, like tribe, became a single, exclusive, politically charged, total and totalising identity only with colonialism.

The colonial state was based on a double discrimination, racial and tribal. Racial discrimination was institutionalised in the central state, and tribal discrimination in the native authority.

Race was said to be about a hierarchy of civilisation, whereas tribe was said to reflect cultural (ethnic) diversity within the native race.

If the central state discriminated against the native race on civilisational grounds, the native authority discriminated in favour of the native tribe on grounds of origin.

Race and tribe are totally modern constructs. Indeed, they are totalitarian constructs, whereby a single identity determined everything: land, governance, law.

The nation-building project in African colonies did not begin with colonialism; it really began at Independence.

The post-colonial challenge was to create a single political community, the nation. To create this nation meant to create a single community ruled by the same set of laws.

This project would require eliminating both sets of discriminations that split the residents of the colony into separate political communities: Race and tribe. It is more or less how we did this with regard to race.

It involved a triple reform with reference to land, governance and law: Removing racially designated land areas; doing away with separate governance for each race; and creating a single law for all, without regard to racial identity.

The trouble was that this reform was partial. It was a reform confined to the elite sector. The reform of the popular sector has yet to be completed.

This is the reform that must do away with discrimination based on the identity we call “tribe.” The conceptual challenge is to distinguish tribe as a political identity from ethnicity as a cultural identity. I want to take three examples to make my point: Nigeria, Congo and Tanzania.

Nigeria went through a brutal civil war in the 1960s. After the civil war, they passed a new “federal” constitution.

It was said that this constitution embodied the lessons of the civil war. The central provision in the Constitution was identified with the clause that said that key federal institutions must reflect the “federal character” of Nigeria.

What were the key institutions? There were three: The civil service; the army; and the federal universities.

What would it mean for these institutions to reflect a federal character?

To do so, each must recruit from all units of the federation — from all states — without discrimination. Each state must have a quota reflecting its share in the total federal population.

The next question was: Who can compete for the quota?

The answer: Only those indigenous to the unit. The final question: Who is indigenous to the unit? The answer: A native, an indigene, only a person born of a father born in that unit.

Let us identify the key consequences of the federal character clause, which was interpreted in such a way that it became an indigenisation clause.

Most obviously, it disenfranchised ethnic minorities in each state, those said to have originated elsewhere. But that did not mean that it empowered ethnic majorities. In reality, it divided the population between those most mobile and those least mobile.

The constitution disenfranchised those most mobile, whether they were poor — like landless peasants, jobless workers, and itinerant traders — or well-off, like merchants, professionals an industrialists.

At the same time, it privileged those tied to the land, in this case too both poor and rich, such as landholding peasants and landlords. It also disenfranchised women who tended to move residence from their natal community when they got married.

The result was a growing structural tension between the economic system and the political system.

Whereas the market economy mobilised more and more strata of the population, moving them beyond local boundaries into a national mix, the political system disenfranchised those most mobile.

In the process, Nigerians found themselves with two different citizenship affiliations, one tribal and the other national. The local affiliation was the more meaningful.

Over the past four decades, the politics of indigeneity has spread in Nigeria. It is not just the population of a state, but also that of a village, that is now divided between indigenes and non-indigenes.

My second example is from Congo. In 1997, a colleague and I were asked to go to Kivu and write a report on the growing political conflict between the Kinyarwanda-speaking population and indigenous Congolese. We toured the full length of Kivu.

We found that there were three main Kinyarwanda-speaking communities in Kivu, in Rutshuru (North), Masisi (Central) and Mulenge (South).

We noted that whereas the Banyarutshuru were considered Congolese, the Banyamasisi were not, and there were divided views on whether the Banyamulenge were citizens or not.

Why the difference? It did not stem from their culture, from the language they spoke, but from when these groups had moved to the Congolese side of the border.

If they had moved before Belgium demarcated the boundaries of contemporary Congo, they were considered indigenous Congolese; but if they moved after, they were considered non-indigenous.

The issue became volatile after independence, especially as a democratic opposition to Mobutu began to organise.

It is worth noting that Mobutu, the dictator, announced in 1972 that all Rwandese who had emigrated to Congo before Rwanda’s independence in 1959 would be considered Congolese, but the Sovereign National Conference meeting in Kisangani in the 1990s considered them foreigners.

Whereas the dictator wooed the minority, the majority disenfranchised the minority — democratically.

The consequence of this decision has been monumental. If you want to understand the causes and issues that have driven the Congo war that is said to have taken the lives of five to six million people over the past decade, you need to begin with the controversy around the citizenship claims of the Banyarwanda — the Banyamulenge in particular — to understand the internal split that has defined the two sides in the civil war, and only then move to how regional powers like Uganda and Rwanda have attached themselves to different sides in the ongoing civil war.

The conflict that led to the civil war in the Ivory Coast was not that different from the Congolese conflict.

The migrants in Ivory Coast came from Burkina Faso in the colonial period.

They too were labour migrants brought to work in the coffee plantations in the north.

After independence, when they demanded citizenship rights, the dictator Houphouet-Boigny agreed, but the democratic opposition in the south did not. Thus were sown the seeds of the subsequent civil war.

What is the lesson? The first lesson is to recognise that these issues stem from how rights were defined in the colonial period.

Should political identity be defined as distinct from cultural identity and therefore based on locality as opposed to origin?

Should the claim to rights be based on individual citizenship or on group identity?

I suggest that the beginning of wisdom is the recognition that there is no one side that is absolutely right and another that is absolutely wrong.

There is a second lesson. To bring it out fully, however, we need to look at another example: that of mainland Tanzania under Mwalimu Nyerere.

It is unfortunate that Mwalimu Nyerere is known more for his economic thought — ujamaa — than for his political practice. I believe that Mwalimu Nyerere’s greatest contribution was as a statesman who decolonised the state, but without using any violence.

It used to be the ABC of Marxist intellectuals at the University of Dar es Salaam that the state is an armed power, and that to dismantle the state you need to smash it, and that you can smash armed power only with an armed force. Thus the romance with armed struggle as the prerequisite to liberation.

My analysis suggests otherwise; more than an armed power, the colonial state was a legal-administrative power.

You could replace the colonial army and police, but you would still not decolonise the state. To do so, you would need to dismantle the legal-administrative power that oversaw a double discrimination: Racial and tribal.

Nyerere's contribution was that he went beyond racial discrimination to remove the institutional basis of tribal discrimination.

First, he dismantled the dual legacy of tribal homelands and native administration, by giving primacy to building a centralised state bureaucracy.

That reform, carried out in the 1970s, was ironically called “decentralisation” because it extended the reach of the state bureaucracy from the centre to the regions and districts.

The details of the reform were finalized by an American consultancy firm called Mackenzie & Associates. This was state-building, not democratisation.

Second, Mwalimu did away with the colonial legacy of a plural legal system based on a split between civil law and different sets of tribally identified customary laws. Instead, he created a single body of substantive law for all citizens, enforced by a single hierarchy of courts.

The consequence of this set of reforms is also monumental: Mainland Tanzania is the only state in this region that has not victimised a group as a group, either as a race or as a tribe. I believe this is because Mwalimu succeeded in creating a single citizenship in mainland Tanzania, regardless or race or tribe.

What is the lesson here? Democracy is not a magic cure so that it can by itself provide a sufficient guarantee of appropriate statecraft, what we now call governance.

The democratic opposition in Congo, as in Ivory Coast, disenfranchised minorities, and created the ground that sustained a civil war.

In contrast, Nyerere, not a democrat but a nationalist and a Pan-Africanist, created the institutional basis of a common citizenship and rule of law in mainland Tanzania.

Next week: How do we move forward? Why the agents of change have to come from the very political class that is our problem today

Professor Mahmood Mamdani is a Ugandan scholar, known for his seminal works: Citizen and Subject; When Victims Become Killers (on Rwanda); and Good Muslim, Bad Muslim. His latest work, Saviours and Survivors (on Darfur in the Sudan) will be out next March. He has taught at the University of Dar es Salaam, Makerere Univerity and the University of Cape Town, and is currently Herbert Lehman Professor of Government at Columbia University in New York. He was this year voted one of the world’s top 10 public intellectuals by two journals, Foreign Policy (US) and Prospect (UK). He was asked by the Kenya Human Rights Commission to deliver this year’s annual International Human Rights Day lecture in Nairobi on Monday, December 15. The above is an edited excerpt of that lecture