Photo courtesy of The World Bank The starting point of the analysis in this article is to question the conventional approach to studying a state. The conventional approach treats the state as a concrete, self-contained entity that has attained a final status. Much of the effort to promote goals such as economic growth, social development and democracy is based on a notion that states have been formed, and now the task is to focus on promoting these objectives. In contrast, this article looks at states as products of historical processes like any other social phenomena. Therefore, state formation is the more relevant term to use in studying a state. States are formed under certain specific historical conditions. They continuously undergo changes and under certain circumstances can even totally disappear. A cursory glance at the history of the world will show this. In addition, state formation always takes place in a global context, consisting of a system of states and global capitalism. The global system changes over time; this in turn has an impact on the state formation process of individual states. Changes in the global system are determined by actions of the more powerful players in the international system. While the capacity of smaller states to influence changes at the global level is limited, these changes have an impact on smaller states. The earliest efforts within the Marxist tradition explained the state as a product of capitalist development. This economic reductionism was replaced by a notion of the relative autonomy of the state. This has developed further to distinguish between the logic of capital and the logic of state power. Now there are many more studies within the Marxist tradition that focus on the autonomous power of the state. Individual states can be seen as strategic spaces. They have a degree of autonomy from other societal processes and cannot be understood by reducing them to any other feature in society. State formation involves developing mechanisms to control territory and to manage state-society relations. This takes place in a specific historical context, where certain state-society relations become more strategic in the state formation process. These strategic state-society relations can be managed either through coercion or consent. When consent overrides coercion in this process, we have states that have legitimacy in society. These states are strong. When coercion predominates, state security is given priority but it undermines the security of individuals and groups in society. These are weak states. A state needs resources to sustain itself and manage strategically important state-society relations. These constitute the economic security of the state. The resources for this have to be secured within global capitalism. The liberal answer to this is to promote markets and openness to global capitalism. Other ideological currents focus on various forms of state capitalism supported by ideologies of mercantilism and economic nationalism. But managing relations with global capitalism to achieve the economic security of the state is an on-going process. This means there is no guarantee that one can achieve a final status of economic security of the state. When a state has enough resources to manage strategic state-society relations through specific policies and is able to meet other critical needs, we have a strong state. When this fails the state resorts to coercive measures, and this affects the pattern of state expenditure and the state becomes weaker. In the post-colonial state formation of Sri Lanka there have been three strategic state-society relations. These are the relations between a centralised state and ethnic minorities; electoral politics and the political system and how they managed state-society relations; and relations between the state and the Sinhala majority in the context of the politics of capitalist transition. All three strategic state-society relations have resulted in conflicts in the post-colonial state formation of Sri Lanka The foundation for all three were laid during British colonialism. This does not mean we can blame colonialism for all that happened during 75 years of post-colonialism. But an understanding of the colonial foundation is helpful, especially to realise how certain ideas about state formation that began during the colonial period still prevail and hamper the formation of a more legitimate state. The period of British colonialism was an important turning point in the history of Sri Lankan state formation. There were various forms of territorial control in the island before the British took over the island. When the British captured the Kandyan kingdom in 1815, the territory came under a single polity for the first time in several centuries. This allowed the construction of a state that formed the foundation of the post-colonial state. The institutional structure, or what is popularly called the administrative structure, is the means by which the state controls its territory and people. It ensures political decisions made at the centre are implemented throughout the territory. The boundaries drawn to form the institutional structure are used to administer the territory. This institutional structure gets strengthened through a judicial system, structures to collect revenue, and coercive mechanisms to consolidate state power over the territory. These are major steps in the reorganisation of geographical space in state formation. The British amalgamated the maritime provinces and the Kandyan kingdom in 1818. Further reorganisation of the territorial space to establish the new state took place through the recommendations of the Colebrook-Cameron Commission in 1833. A new system of spatial identities replaced what prevailed before the British took over control of the land. Spatial units were named using the points of a compass. The concepts of modern cartography became an instrument of colonialism and were used to get rid of the spatial identities in the Kandyan kingdom. In addition to establishing the institutional power of the state, state formation involves collecting information about the population. This is also a part of the technique of state formation to control territory and manage the population. A periodic census is the foundation for this. The division of the population into identity groups is a critical aspect of the census. At the level of society identities change over time – they are historical constructs. It is not unusual for individuals or groups of individuals to have multiple identities. But when a mechanism of state formation converts these fluid categories in society into divisions that the state recognises, they not only become rigid and static, but also begin to play a role in political struggles to control the state. Race was the term first used to identify these divisions of the population. This was replaced by ethnicity, and it became a major category in the post-colonial state formation of Sri Lanka. In addition, these techniques of state formation began to identify a particular geographic space with a particular ethnic group. A central contradiction of these techniques of state formation was that on one of the techniques created a centralised state that controlled the entire territory. But on the other side another technique not only gave a new meaning to ethnic identities, but also identified territorial units within the territory with different ethnic groups. A single identity, which was called a national identity, was supposed to transcend these contradictions. A centrally controlled state with a single national identity was a transfer of a European idea of state formation to Sri Lanka. However, the formation of this single national identity was not successful. What happened was that the identity of the Sinhala Buddhists becoming the national identity. The notion of a state with a single identity still dominates our political discussions of state formation as if this is the only way to build a state. Questioning this fundamental assumption is an essential condition in building a more legitimate state in a society with multiple identities. Using elections to choose the political elite who control the state was established during British colonialism. The territorial system of electorates was preferred. This added another territorial dimension in state formation. When it came to a voting system, the first-past-the-post system of elections was chosen. The regime that came to power in 1977 replaced this with a proportional system of elections. Areas with the Sinhala majority had more influence in choosing who came to power in both electoral systems. The very first act of defining the citizenship of the post-colonial state disenfranchised the bulk of the Indian Tamil population. According to the 1946 census they formed the numerically largest minority ethnic group. With this step it was clear that ethnicity would be a major factor in electoral politics and the political system. Such a political system could never contribute to transcending ethnic divisions socially or spatially. Research in other parts of the world with a long record of continuous voting, such as Western Europe, shows that electoral politics and political systems are an important factor in constructing a national political space. This never happened in Sri Lankan state formation. Instead, electoral politics and the political system produced regional political spaces with ethnic characteristics. Therefore, it is more apt to discuss election results in Sri Lanka by dividing it into ethnic political spaces rather than treating the state as a unified entity. Even now the discussions on electoral reform do not question these fundamentals of the electoral system that contributed to the failure of state formation. The liberal idea that society is a collection of individuals who are given a free vote still dominates. The essentialist categorisation that Sri Lanka is a democracy also makes it difficult to undertake a historically informed discussion. Sri Lanka had democratic institutions. But politics, which focuses on how power is managed in society, has been far from democratic and peaceful. Some even forget that Sri Lanka did not have universal franchise for a long time after the bulk of the Indian Tamils were disenfranchised. There is a need to get over these assumptions so as to begin a discussion on electoral and democratic reforms to build a more legitimate state. The third important strategic variable in state-society relations are relations between the state and the Sinhala majority in the context of the politics of capitalist transition. Capitalist transition within a state is a process that involves changing institutions or the rules of the game so that markets become the primary mechanism for resource allocation. These changes must be legitimised at an ideological level. When institutions to establish markets are successful, they become ideas that seem to be natural and common sense, thereby creating a hegemony. The establishment of the hegemony of markets is not a technocratic process, but a political process. Conflicts and struggles are always a part of this. The process of capitalist transition takes place in a particular society with its own history. This means that capitalism is not some sort of model. It is shaped by political struggles and historical processes in a particular context. Finally, it takes place within global capitalism. British colonialism was a period when there was an intensive process of capitalist transition. Political unification of the territory facilitated the expansion of colonial capitalism. This was the third process that reorganised space during British colonialism. The plantation economy that developed during British colonialism was not confined to the central part of the country, as sometimes believed. The tea industry transformed the bulk of the highlands, parts of Sabaragamuwa, and the districts of Galle and Matara. Rubber and coconut encompassed the south-west quadrant of the country. What came to be known as minor export crops were in many parts of the island. Data in the 1946 census demonstrates several important spatial outcomes of plantation capitalism. In 1946 districts that at present constitute the Western Province, Sabaragamuwa Province, and Galle, Matara, Kandy, Nuwara Eliya, Kurunegala and Badulla Districts accounted for 95.1 per cent of tea acreage, 92.4 per cent of rubber acreage, and 81.2 per cent of coconut acreage. These areas also accounted for 81.2 per cent of the population. This means that at the time of independence the bulk of the population was linked to the export-oriented plantation economy. People were involved in diverse ways of earning a living such as agriculture, selling their labour, trading, fishing, etc. The notion of characterising Sri Lanka primarily as an agrarian society was an idea that emerged later. This was supported by the dual economy thesis, Sinhala nationalism and research that focused primarily in Kandyan areas. The picture was very different in the entire wet zone where the bulk of the population lived. The socio-economic impact of capitalism is always unequal. Some sections of the Sinhala majority benefitted more than others from capitalist transition. In other words, although the Sinhalese were unified in ethnic terms, they were divided in class terms. In a state where the state was identified with the Sinhala majority, and a political system where the Sinhala majority were the deciding factor in who came to power through elections, the political elite who controlled the state could not ignore the distribution of unequal benefits through capitalist transition. The inequality generated by capitalist transition within the Sinhala majority could always combine with the Sinhala nationalism that legitimised the state to oppose the regime in power. The opposition to regimes could also turn into opposition to capitalism, and a general opposition to the state itself. In order to meet this challenge, the post-colonial state developed a number of policies. The particular spatial distribution of the population at the end of the colonial period and the impact of economic recession in the 1930s had a lot to do with specific policies that were developed. These are what are usually called welfare policies. The use of the term welfare ignores their strategic role as a technique of state formation. This makes it easier to argue that these policies are luxuries we cannot afford. The ability of the state to continue with these policies depends on the performance of the Sri Lankan economy within global capitalism. Economic policies that managed relations with global capitalism varied depending on the ideological orientation of the regime in power. While one section of the political elite preferred policies that gave prominence to the private sector, markets and openness to global capitalism, another section depended to a greater degree on the state playing a bigger role in the economy. This was backed by various ideologies including economic nationalism. In post-colonial history there were several instances when the strategies of managing relations with the Sinhala majority in a context of capitalist transition broke down. This resulted in protests, sometimes violent challenges to the state, and state repression. We have just witnessed the latest episode of such a breakdown with familiar outcomes. This is an area that demands a new discussion on social policy with a focus on inequality. Looking at the state as an entity that is being formed helps to identifying strategic state-society relations in state formation. This process in turn provides a framework for a more comprehensive approach for discussing policy issues that Sri Lanka is facing today.