Nation, Ethnicity, Milieus, and Multiple “We’s”. The Case of Kenya

Abstract

The title of the volume “Future Africa—beyond the nation?” has several implications. Nation is presented as an entity relevant to identification and identity; and in the combination with “future”, nation implies a political vision. It is not hard to find good examples in respect of these implications. However, there are other entities important for to political identification. Often, they do not go beyond the nation but refer to smaller collective identities, such as ethnicity. The revived debate on “the middle class” implies that particular social groupings, such as class, may play a role, too. The question is how relevant are the nation and other collective political identities in Africa, and are they exclusive? Looking at the case of Kenya, we see on the one hand that collective (political) identities, such as ethnicity, are mobilized especially during elections. On the other hand, these collective identities are less dominant in everyday life and give way to different conducts of life (conceptualized as “milieus”) that are less politicized. We see people maneuvering between multiple “we’s”. Strong political identities are mobilized only in particular conflict-loaded situations that restructure identities in simple binary oppositions of “we” and “they”.

1. Introduction

In the discussion on politics and society, the so-called ‘nation state’ is a general point of reference. First, this reflects the dominant pattern of international politics, in which, despite globalization, the nation state is still a crucial entity that enacts transnational agreements and is the main level for policy decisions. At the same time, despite all inequalities at the national level, nationality defines individual opportunities in the international system with regard to access to education, freedom to travel, or economic chances. As we learn from neo-instituionalist scholars, to gain international legitimacy in the world of ‘nation states’, political entities need to be identified as nation states (Meyer et al. 1997). Second, the nation state and its ‘society’ are still crucial categories in social science research. International statistical data provided by the World Bank, UNDP, OECD, and other multilateral institutions follow the nation state principle. Theories of society and political systems or democracy refer to the nation state as a quasi-given category.1 The title of this special issue, ‘Future Africa—beyond the nation?’, follows this line of thought by asking whether there is something that goes beyond the nation state. Together with the notion of a future and with implications such as jihadism, confederacies, and pan-Africanism, nation is linked with political visions, and the nation or a wider setting act as entities creating political identity. In our search for political visions and identities ‘beyond the nation’, we may start by asking whether there are national identities in Africa that might be overcome. If we refer to democracy theory, it is obvious that political mobilization inside a society cannot solely refer to national identity. The typical mobilizing sub-national identities are class, sometimes also ethnicity or religion, and even region.

However, at the same time, nation is a relevant category for identity. One obvious case in Africa is the national political mobilization in connection with independence. Other cases are wars of secession, such as Katanga or Biafra. Sometimes, national identities go beyond borders. This was the case with the idea of greater Somalia or greater Morocco and with the ongoing Tuareg conflict with a Tuareg identity that crosses national borders between Mali, Niger, Algeria, and Burkina Faso. National identity is strengthened by wars between African states. These are extremely rare, but may happen, as in the case of the war between Ethiopia and Eritrea (after a peaceful secession). Nationalism also increases in conflicts circling around prejudices against immigrants. Typical cases are the attacks against Mozambican workers in South Africa, and the less internationally known conflict with Somalians in Kenya. National identities are obviously important in conflicts between nations or relating to national issues. However, can we find expressions of national identity without fundamental conflicts? Typical events that raise national feelings are international sport activities. For instance, when a Nigerian football team plays in the World Cup, it can count on the support of all Nigerians. In a similar way, Kenyan runners who take part in world championships or the Olympic Games will be supported by all Kenyans, even though they may come from a particular region, such as the Rift Valley area, and are members of particular ethnic groups, such as the Kalenjin (often from the Nandi subgroup). They will nevertheless be supported by members of other ethnic groups, such as the Kikuyu or Luo. Other cases where nationality plays a role are comparisons between countries with regard to the economy, conditions of life, or corruption. Also, if one asks African students in Europe or North America where they come from, they will name their home country, not the city or district they see as their real home. The simple conclusion is that there are national identities in Africa even in the absence of fundamental conflicts. However, national identities are stronger and much more outspoken in cases of conflict and in situations where inclusion or exclusion plays a crucial role. In situations without any reference to the nation as a source of political and cultural identity, people’s identification with the nation is much weaker.2 Nation is not the only entity of (political) identification: there are multiple we’s that play a significant role in everyday life. These include religion, clan, class, genealogies, race, or origin (ethnicity). In this paper, I will ask how these multiple we’s and national identity relate to each other with regard to the example of Kenya.3

Why Kenya? Kenya is a country with few international conflicts, and even its involvement in the Somali conflict does not interfere too much with everyday life. There are terrorist attacks, there has been political violence, nevertheless Kenya is one of those countries with a fairly stable political system. All power changes in Kenya have followed the constitution. A second reason why Kenya is an interesting case is that we can identify visible subgroups that represent multiple we’s. This is most obvious at election times. Conflicts between ethnic or regional factions are common during elections and led after the 2007 election to a fierce conflict that brought Kenya to the edge of civil war.4 Thus in a first step, we will look at political identities, political visions for the future, with regard to the elections. In a second step, we will look for identities and subgroups that are significant in everyday life and their relation to political visions. This will lead us in a third step to discuss some misunderstandings with regard to common political identities. The paper will end with conclusions regarding the question when we-groups are relevant.

The empirical data is drawn from the research project ‘Middle class on the rise’ which was a part of the larger project ‘Future Africa—visions in time’ funded by the German Federal Ministry of Education and Research (BMBF) at the University of Bayreuth (2012–2018).5

2. Political Identity in Multi-Party Elections in Kenya

Political identity in Kenya is most obvious during multi-party elections. Kenya is one of those countries in Africa in which regular multi-party elections have taken place since the early 1990s. With regard to the criteria of democracy theory, they were successful and led to changes of power. The notion of multi-party democracy is clearly influenced by European, and to a lesser extent North American, experience, and leads to the assumption that multi-party democracy organizes competition between different parties with different political programmes. The assumption behind this understanding is that people with different, more or less clearly defined interests are represented by different parties. This system worked very well in the “short” 20th century in Europe. The background for this system was a class society. Workers related to the left-wing socialist or social democratic parties, while the conservative (petty) bourgeoisie and conservative peasants supported the conservative parties. In countries with proportional representation, smaller parties represented other groups, for example liberal parties representing liberal professionals. These class-based multi-party systems might never have been as simple as described here. However, at least since the 1980s they have lost their clear shape. In countries with proportional representation, new parties have emerged, such as “green” parties that do not follow this class pattern, while in majority vote systems the clear-cut difference between bourgeois conservatives and workers’ left-wing parties has become eroded. When multi-party democracy was promoted for Africa, somehow this notion of a class-based system was in the background.6 At least there was an assumption that political parties should represent different programmes and political aims. How does this correspond with multi-party elections in Kenya?

If we look at the presidential elections of 2017, we see that the different parties are not easy to identify via their programmes. The programmes are quite general and do not differ very much. The parties have no clear political and programmatic profile. If we look at the map showing the distribution of the votes, we see a regional distribution for the two leading candidates. If we compare this map with a map of ethnic groups, it becomes obvious that the preference for one or the other presidential candidate roughly follows the regional distribution of ethnic groups. The map shows regional-ethnic blocs that competed against each other in the presidential election. The data for the parliamentary elections follow a similar pattern. It is clear that in the 2017 elections, political mobilization in Kenya still followed broadly ethnic patterns. This is not surprising, because Kenya is one of those countries where ethnic politics are well established and have been a feature of multi-party elections since 1993. Ethnicity also played a decisive role in the one-party system, which was marked by an arrangement of regional and ethnic patron–client pyramids (Barkan and Okumu 1978; Cheeseman 2011; Hulterström 2007; Tamarkin 1978). This pattern is strengthened by the fact that in large parts of Kenya one ethnic group still dominates. These ethnic groups claim their particular part of Kenya as their ‘home area’. Only those parts of the Rift Valley that were occupied by white settlers (the former ‘White Highlands’) and the coast are ethnically more heterogeneous due to in-migration. This of course also applies to the larger cities.

The observation that it is impossible to link the Kenyan parties with political aims that point at particular class interests underlines the fact that class does not play a decisive role in political competition. An issue that showed differences between the parties in the 2000s was the political power structure (centralized or decentralized, more or less power for the president). Even this discussion was fuelled by the interest of politically weaker regional-ethnic blocs in gaining a certain political autonomy. Another political issue was the regulation of abortion (we will come back to this topic later). Both points were at stake when the new constitution was debated and finally approved via popular vote with a certain degree of decentralization (Murunga et al. 2014). The question is still why there are no class-based parties.

There is a simple explanation: there are no clearly divided classes in Kenya. Despite obvious inequality, people cannot be structured along class lines. Even though it is not the main point of this article, this needs at least a short explanation (see Neubert 2016, 2019b). Marxist and Weberian class concepts link class either with modes of production or with occupation. What we find in Kenya, as in many other African countries, too, are families that include peasants, blue- or white-collar workers, and self-employed people at the same time. In a family, the mother might be working on the family plot, one of the sons might be a factory worker, the well-educated daughter might have a well-paid white-collar job in a service organization, and another son might be self-employed in the so-called ‘informal sector’. Often the same person combines different sources of income: a person may be employed, own a small business, and have a farm, either for subsistence or for commercial production. People with urban income may invest in agriculture. If we look at people’s life courses, we will see that people often change their occupational position, moving from being self-employed to being a farmer, or changing between white- and blue-collar jobs. Thus, we observe a biographical alternation of occupational positions. This combination of modes of production was first described for low-income groups in the 1980s (Elwert et al. 1983; Smith et al. 1984). However, on a second look, we see that it also holds true for middle- and higher-income families.

This is not a new pattern, as especially for middle-income groups in Kenya, it developed already during colonial times. Before colonization, the ethnic groups7 in the area that later became Kenya were mostly acephalous. Power was in the hands of the male elders and social roles were assigned according to age and gender. There was no stratification based on income or economic assets. Exceptions to this were the few coastal towns, which had a more differentiated social system with differences of wealth, power, and status. With colonization, a state structure and capitalism were introduced into Kenya. This had consequences for the social structure, as in other African countries. Due to economic changes, blue-collar workers were needed, for instance in the harbor, and later also for the railways, while the small African minority that had access to Western education took white-collar jobs in the administration and the still small private sector. At first, primary school was sufficient; later, secondary school certificates became important for access to attractive positions. This educated group formed the so-called ‘African elite’, in fact a new middle-income group, in colonial society. The very few university graduates had access to higher positions. With the colonial ‘Development & Welfare Acts’ in the 1940s, the improvement and commercialisation of agriculture received support through the Swynnerton plan (Hailey 1957, pp. 203, 1323; McWilliam 1976, pp. 262–64). This included a process of land registration that started in the most densely populated areas. Thus, land that in pre-colonial times was communal land administrated by the elders moved into private hands and could be sold and acquired in freehold. A study by Gavin Kitching ‘Class and economic change in Kenya: The making of an African petite bourgeoisie 1905–1970’ (Kitching 1980) analyzes the process of social change linked to these developments. In a nutshell, Kitching shows that urban middle-income earners had farms and invested in commercial agriculture. Those with sufficient capital acquired additional land. The success of Kenyan agriculture was connected with urban employment. These middle-income families combined their income from employment with farming, and some also started small businesses. The group addressed as ‘petite bourgeoisie’ by Kitching combined different sources of income and were at the same time peasants, workers, and (small) entrepreneurs. Another study by Kongstad and Mönsted (1980) came to a similar conclusion. This pattern is still found. Githinji’s study on inequality, with a large sample of rural households in Kenya, confirms the combination of farm and off-farm income for the time after independence (Githinji 2000). The interview data from our project underlines the ongoing pattern of the combination of different sources of income.

There is a second reason why conventional class concepts cannot be applied in Kenya. Class concepts imply a relatively stable status (Weber) or class position (Marx). It goes without saying that this refers to the family. Social science contributions to the ongoing debate on the African ‘middle-class’ point to an important fact. They challenge the notion of a stable middle-class position and describe the uncertainty and insecurity of the middle-income group (Benin: Alber 2016; Heilbrunn 2014; Ghana: Hamidu 2014, p. 187; South Africa: Kriege 2015, p. 112). This is also a clear finding of our Kenyan study.8 People face the risk of falling to a lower position, or even back to poverty; upward and downward mobility is part of the biographical experience of many people. The statistical growth of the middle-income group shows a positive balance in terms of those who move up and those who move down, and hides the overall volatility of the socio-economic position of middle-come earners. Because of this uncertainty, and the limited formal social security provisions, semi-formal and informal networks are crucial to the security of families. People expect to support poorer members of their extended family, or fellow members of other networks (e.g., church, neighbourhood, local community, rotating saving groups), while at the same time they can expect to receive support from better-off network members. Again this is a well-established pattern that has existed for a long time and is still important today.9 These patterns of support systematically bridge different socio-economic positions. The knowledge that a shock might lead to a situation where a current giver needs support strengthens the feeling of belonging. There are groups that try more or less successfully to escape from these obligations, but for large parts of the society, this link across socio-economic positions is part and parcel of people’s lives. This system can be a burden for the better-off members of the network, and is often accompanied by conflicts over whom to give how much and when, but it is an accepted social obligation and offers at the same time a basic safety-net. As a result, even those with better positions share the interests of poorer groups, either because they are closely tied to them, or because they realize that they might fall into a similar position.

Against this background, a party that promises to support peasants, another party that promises to support laborers, and a third party that supports the interests of those who are self-employed, will not have any clear constituency, because the potential electorate has all these different interests at the same time. Similarly, parties distinguishing between the ‘poor’ and the middle-income earners would not reflect the existing structural interlinkages.10 Thus, another pattern of political mobilization is much more successful. As long as political competition is linked to the development of certain regions in terms of infrastructure or economic support, the parties that follow a pattern of regional blocs and overlap with ethnicity seem to be much more rational than class parties.

3. Everyday Identities in Kenya

If we look at the election results, we might draw the conclusion that identity in Kenya follows ethnicity. Everyday experiences seem to support this idea. Ethnic affiliation obviously plays a role in everyday life. For instance, when people mention their names, their interlocutors will usually try to identify their ethnic affiliation from the name; and very often, they will be right. This awareness of ethnic identity might not have direct consequences for the interaction or the relationship, but it is there. Particular football clubs are identified with ethnic groups: ‘Gor Mahia’ is seen as a Luo club, ‘Leopards’ as Luhya, and recently ‘Mount Kenya United’ as Kikuyu, Embu, and Meru.11 In the public discussion, there is at least one group that is very outspoken with regard to political rights for ethnic groups. One of the issues which is politicized along ethnic lines is access to land understood as ‘ancestors’ land’. Thus, land conflicts are often ethnic conflicts, or at least have an ethnic undercurrent.

However, this focus on ethnicity presents a visible surface, but no more than that. Ethnicity does not represent all relevant socio-cultural identities, or multiple we’s in the sense of identification with a certain group. While ethnicity is certainly present in everyday life, people’s occupations and their different conducts of life and lifestyles are at least as important as their ethnic affiliation. The way people live, their values, their preferences, do not systematically follow ethnicity.

In societies with clear class differences, lifestyles and conducts of life are linked to class. When neither class nor ethnicity determines lifestyles and conducts of life, we need a new approach to capture socio-cultural differences. For this purpose, we can refer to a sociological concept developed in Germany in the late 1980s. At that time, sociologists identified an ongoing differentiation within German society which they could not capture using simple class terms. Thus, two concepts gained importance: ‘lifestyle’ and ‘milieu’. Lifestyle referred mainly to everyday practice, while milieu combined everyday practice with norms, values, and attitudes. With regard to our question of identification with certain groups, the milieu concept is more promising, because it includes values. I will not revisit the still ongoing discussion on lifestyle and milieu12. I want to show that the milieu concept offers a valuable starting point to analyse the multiple patterns of identification in Kenya (Neubert 2019b, chp. 7; Neubert and Stoll 2015, 2018; Stoll 2017, 2018).13

Milieus are “… sub-cultural entities inside a society that capture people with a similar view of life and way of life …” (Flaig et al. 1993, p. 55, my translation). The term “sub”-culture refers to a cultural unit inside a society that represents a certain number of people in the society. This does not exclude the possibility that people belonging to a certain subculture might identify themselves with people in other societies or in other nations. Especially in this age of globalization, subcultures may be trans-national14. Milieu thus describes a certain affiliation to a particular way of life and a particular set of values. The advantage of this concept is that it does not pre-define certain milieus. The identification of different milieus is an empirical question to be answered for each society anew. Flaig, Meyer and Ueltzhöffer (Flaig et al. 1993, p. 71) propose a set of buildings blocks that help us to ask the right questions.15 This was initially helpful for analyzing socio-cultural differentiation in Kenya, but we needed some adaptions and we added some points.16

Based on our empirical data, we developed a framework adapted to the Kenyan context. We identified six main socio-cultural dimensions that mark differences between the milieus (Neubert 2019b, chp. 8). The first of these is ‘advancement’, which refers to motivation for upward mobility, career, and education. Nearly every Kenyan will say that education and upward mobility are the main goals in their life. However, the intensity of investment in advancement with regard to time and resources varies considerably. The next dimension is ‘moral values’, ranging from liberal values to very conservative values and gender roles. ‘Rural links’ describes a person’s bonds with their rural home,17 ranging from disinterest in the rural home to the rural home as identity marker and (main) reference point. ‘Consumption’ refers to the amount of resources spent on consumption rather than education and advancement, or investment in business or agriculture. ‘Political activity’ describes whether a person is politically silent, or politically outspoken in public, taking part in political actions and expressing political ideologies or future visions. People with limited political activity may, and regularly do, have a political position and political opinions, but they are not outspoken, they do not express these opinions actively, and thus they differ from those who join in, or at least support, public political activities. The last dimension is the ‘social scope of future aspirations’. This refers to the entity or group that is included in one’s vision of the future. Whose future do people have in mind when they express their visions of the future? Do future visions and/or expectations for the future refer to society as a whole, to a large group, a small community, the extended family, the nuclear family, or just to the single individual, ‘me’?

With these six dimensions, we describe different milieus that may overlap in one or more dimensions, but are different in other dimensions. However, not everybody in a particular milieu lives exactly the same life. There are individual differences. In addition, some people may identify completely with the elements of a certain milieu, while others comply with only some elements, or live between different milieus. Thus, what we describe are ‘milieu cores’, as a kind of ideal type that never corresponds completely with reality. We should also be aware that people may change their milieu during their life course.

With regard to our question of multiple we’s and their relation to the nation and political visions, two dimensions are of particular importance: the ‘social scope of future aspirations’ and ‘political activity’. As we will see, the milieu concept helps us to identify differences and thus different we’s. Based on our research in Kenya, we identify seven milieus, most visible in urban contexts.18 There may be more milieus but these seven offer a good overview of the different conducts of live and values.

The ‘neo-traditionalists’ are a politically visible milieu. They do not have a dominant career orientation. They have conservative values linked to ethnic customs and traditional gender roles. They maintain close bonds with their rural home and lay emphasis on rural investment in their home area. In town, their networks are mostly mono-ethnic and they underline the importance of their local language. Their consumption patterns are not very distinct, but they usually have no problem with alcohol and partying. They can be mobilized for ethnic politics and there is a potential for micro-nationalism. The social scope of their future aspirations tends to be their ethnic group, and this might be combined with a notion of Kenya as a federation of ethnic groups. In line with ethnic custom, the extended family plays an important role. This neo-traditional milieu is socio-economically heterogeneous and includes people in low and high positions.

Another political outspoken milieu is the ‘liberal cosmopolitans’, but their orientation is very different from that of the neo-traditionalists. The ‘liberal cosmopolitans’ are career-oriented and keen to invest in education. They pursue liberal cosmopolitan values, and they are politically committed with regard to human rights, democracy, and gender equality, may be also including acceptance of gay sexuality; they support ecological sustainability and work for economic and social development. Their networks are multi-ethnic, often with strong professional and/or private links overseas. Consumption is part of their lifestyle, but they avoid showing off. They can afford the usual set of consumer goods, such as a well-functioning car, good clothes, maybe fashionable, but not too expensive. Usually, they are still part of their extended family networks and are interested in the well-being of their rural kin. Often they support their home region or their home village, fired by the notion of ‘bringing development’. The social scope of their future aspirations ranges from the individual level to the societal level. People in this milieu tend to be in middle and upper range positions. It consists mainly of the staff members and constituencies of NGOs. However, not every employee of an NGO is part of this milieu.

An important group of milieus are the committed religious milieus. We focused in our research on the relatively large Christian milieus knowing that there are also smaller Muslim or Hindu milieus. Not every religious person is a member of a committed religious milieu. With committed Christian milieus we mean people whose life centres around the church and is closely linked with the church community and different church groups. This would apply in a similar way to Muslim and Hindu milieus. We speak of Christian milieus in the plural because there are significant differences between at least two milieus.19 These are committed conservative Christians, mostly members of the established so-called ‘colonial’ churches (e.g., Catholic, Anglican, Lutheran, Presbyterian), and committed Pentecostal Christians. Beside their strong links to their church community, they share conservative morals and a belief in traditional gender roles. Their social network is focused on the church community (mono-ethnic or multi-ethnic depending on the type of church) with a certain responsibility for other members of the church community. They are mostly politically invisible. However, they might speak out on moral issues, such as an explicit refusal of homosexuality or the fight against abortion. They tend to strictly refrain from drinking alcohol and from going to places that serve alcohol. Besides these commonalities, there are significant differences. Committed conservative Christians consume modestly according to their level of income and refrain from showing off. While they are career-oriented and aim at advancement, this is not always a top priority. Committed conservative Christians maintain links with their extended family and their rural home, but without any particular emphasis. The social scope of their future aspirations includes the nuclear family, the extended family and rural home, and the church community. This milieu ranges from lower to top socio-economic positions. For committed Pentecostal Christians, career advancement is the top priority. They display success, with conspicuous consumption. The church requires them to cut their family bonds, with the result that the church community and the nuclear family constitute the social scope of their future aspirations. Rural links no longer play a role. Committed Pentecostal Christians may be found in a wide range of socio-economic positions, but tend to be in the middle of society.

There are some commonalities between the milieu of ‘social climbers’ and the committed Pentecostal Christians. Career and advancement are top priorities and they try to limit, or even cut, family connections and links to the rural home. They are politically invisible and do not actively intervene in public debates. In contrast to the committed Pentecostals, they refrain from showing off, consume only carefully, and give a high priority to reinvesting their income. Their values are much more liberal, with notions of more equal gender roles. Their main network is the nuclear family, and membership of wider networks (or the church community) is played down. Thus, the social scope of their future aspirations is just the nuclear family. Social climbers are found in lower and middle range positions.

The milieu of the ‘stability-oriented pragmatics’ are not visible as a milieu in public. Their life centres on the family. This includes the extended family in the rural home area. Family visits are a regular activity. They work hard, mainly with the goal of maintaining their social position. Thus they avoid risks. They enjoy consumption according to their income, and their investment in advancement or in enterprises or in agriculture is limited. They have relatively conservative values with conservative gender roles. They are politically invisible. Generally, the social scope of their future aspirations is the extended family. Stability-oriented pragmatics are mostly low- and lower-middle-income earners, but they can also be found among the better-off middle-income earners.

The last milieu to be presented is that of the ‘young professionals’ (already described by Spronk 2012). They are a very small and a particularly urban group. They have a very good education and are in upper-middle and high positions. Young professionals are young, from the twenties to mid-thirties. They are usually either not married or have no children. They have liberal values with equal gender roles. They are clearly career-oriented and hardworking and invest in business. Due to their good salary, they can often consume in a conspicuous way. They like to dine out and go clubbing and to show their success. Their social networks are multi-ethnic and they are not politically active or publicly politically outspoken. The social scope of their future aspirations is mainly individual, maybe a partner, but not the family or a community.20

Overall, we see that only two milieus, the neo-traditional and the cosmopolitan liberals, are politically outspoken and have a political vision that transcends the family and the community. Their we’s relate to the nation. In the case of the liberal cosmopolitans, the Kenyan nation is part of the social scope of their future aspiration. The neo-traditionalists’ political vision centres on the ethnic community with a potential micro-nationalism. Thus, the ethnic group has a status similar to that of a ‘nation’. There are hardly any separatists in Kenya. However, neo-traditionalists want respect for their ethnic rights and traditions; indirectly, they think that Kenya can provide a frame for the different ethnic groups. Apart from the neo-traditionalists and liberal cosmopolitans, there is no group with future aspirations linked to a particular political agenda or with Kenya as nation. Belonging to Kenya is taken for granted, but has no real consequences, either in everyday life or in future aspirations. Except for the two politically outspoken milieus, the social scope of future aspirations and people’s we-groups are mainly limited to the nuclear and the extended family; in some cases, they refer to the church community, or in the case of the young professionals, just to an individual ‘me’. Wider future aspirations and we-groups with reference to larger parts of society are an exception.

4. Misunderstanding of Common Political Identities

Theories of multi-party democracy at least indirectly link political identity either to class or, as in the case of Kenya, to ethnicity. In this reading, identity is always linked to political visions and to larger groups. This might be a nation or an ethnic group or class. At least in Kenya, this kind of political and national identity is of limited significance between elections and in the absence of larger conflicts. In Kenya, identification with certain values and conduct of life follows a much more differentiated pattern, that of the different milieus. Even more important is that people’s everyday visions are usually not political, and political identity does not play an important role. There are subnational identities, and only two particular groups, the neo-traditionalists and the liberal cosmopolitans, follow the idea of a group linked with a political vision and larger group identity. However, as elections in Kenya show, ‘regional-ethnic’ we’s may be successfully mobilized. Different we’s can be observed over time, the political regional-ethnic we appearing in connection with elections. The same we may lose its political implications in everyday life but is still identifiable through people’s names or home area or favorite football club. When it comes to visions of the future or future aspirations, except for the neo-traditionalists and the liberal cosmopolitans, much more restricted we’s are the points of reference, the church community, the extended family, the nuclear family, or, in the case of young professionals, ‘me’. These smaller we’s do not have a clear political connotation. Our empirical study of Kenya shows that different we’s are mobilized in different settings and situations. The question is not only under which circumstances political identities become relevant, obviously in case of conflicts, but how the transition between different we’s take place.

The post-election conflicts in 2007/8 are the most significant event in recent Kenyan history. Two ethnic blocs clashed over the counting of votes. The ruling party led by Mwai Kibaki, a coalition of Kikuyu, Embu, Meru, and some other groups, was challenged by the opposition party supported by the Luo and Kalenjin communities and other ethnic groups. Both parties claimed victory and the opposition candidate Raila Odinga accused the ruling party of rigging the election. After ongoing mass demonstrations, the conflict escalated and violent fighting started, especially in the Rift Valley and in parts of Nairobi. Kenya was on the brink of civil war. The country was divided into two large factions. Although a high number of people died (more than 1000), security forces and the army prevented a full-fledged civil war. In the end, a compromise was negotiated and the two factions formed a coalition government. Neither the reasons for the conflict and the question of guilt, nor the following political quarrels, can be discussed here. They have already been the subject of intensive research.21 The point of interest is the different we’s in this process. At election times, multi-ethnic churches, multi-ethnic associations, or NGOs face internal tension because of different the political positions of their members; but this usually does not threaten the association as such. In the case of the post-election violence in 2007/8, the conflict became more and more intense, to the point that people had to choose one side or the other. This led to a rupture inside multi-ethnic associations and of course in society as a whole (Daniel and Neubert 2014). The post-election violence started with political dissent, developing into political protest and political violence. Even though only a minority was involved in violent clashes, at least at the beginning, they could count on the sympathy of their supporters. There were two turning points in the process. The probably justified accusation of rigged elections led to a ‘moment of intransigence’ that divided society into two parts.22 Moments of intransigence are situations where a conflict issue dominates all other debates. People have to choose one side. All those who do not identify with one of the conflicting groups are seen as the enemies of all the others. The multiple we’s described above lose their relevance as a means of identification. The debate and the conflict is about ‘we’ and ‘they’ (Neubert 2004). The rhetoric of the leaders of the two camps fuelled this moment of intransigence. At the same time, the post-election violence shows that such a division does not automatically end in civil war, but may be bridged and the violence can be stopped. Besides the intervention of security forces, which brought the escalation of violence to a halt, another process gained importance. Members of civil society from both camps were horrified by the atrocities and the high number of evicted people. They realized that the conflict had escalated up to a point where it could no longer be controlled. Church leaders and journalists started to demand an end to the violence, regardless of whether the elections were rigged or not. This led to calls for peace by the foreign embassies and other expatriates, and by some of the mass media. They started a discourse in which the politicians were blamed for dividing the Kenyan nation. This promoted the idea of a national identity that excludes political conflicts leading to civil war. People’s subnational identities were exchanged for the one national identity. Before this event, a national political identity did not play an important role in Kenya. Thus, calls for peace in the nation overcame intransigence. This led to the solution of a coalition government. Once this was in place, the underlying conflict came to an end or at least to a halt.

One result of the political compromise was the drafting of a new constitution. One disputed issue in the constitution (besides the decentralisation question) was a revision of the radical prohibition of abortion23. This led to a political conflict that cut through the regional-ethnic blocs formed during the post-election violence. A coalition of conservative Kenyans rejected the constitution because of this reformed abortion paragraph. This coalition, led by religious and neo-traditional leaders, integrated the different committed religious milieus and the neo-traditionals. The pro-constitution movement was led by cosmopolitan liberals, who succeeded in the end when the constitution was accepted by popular vote. Even though this political conflict was less fierce and non-violent, it showed that different coalitions can be formed in respect of different issues. In this case, a conservative ‘we’ stood against a liberal ‘we’. This underlines that the value-based milieus represent different we’s that cut across regional-ethnic and socio-economic differences.

These examples show how different entities of identification, multiple we’s, may change during a political conflict and between different political conflicts. In the post-election violence, the milieus lost their importance and ethnic identity gained power up to the point of creating ‘we’ and ‘they’. With the introduction of the discourse of the Kenyan nation, ethnicity was painted as a threat to Kenya in general, and radical neo-traditionalists became the minority against the rest of the Kenyans. With regard to the abortion question, different camps developed whose main protagonists represented different milieus. Political identities constitute just one particular we that may change according to the political issue, and when no crucial political issues are at stake political identity may lose its importance.

5. Conclusions

The empirical information presented here relates mainly to Kenya. Is Kenya a special case? Of course, Kenya has a particular history and a particular social and political development. It is one of the few countries in Africa south of the Sahara, which was under the influence of white settlers during colonisation, even though their number remained small. Current land conflicts, especially in the Rift Valley, are a consequence of the settlement of colonial farmers in the so-called ‘White Highlands’. After independence, this area was re-opened for African settlement. Since then, different land claims have collided and set the background for fierce ethnic conflicts. As mentioned at the beginning of this article, Kenya is one of those examples in Africa where ethnicity is a decisive factor in politics.24 However, there are also general conclusions possible. These are worth being researched in other settings and countries, too.

The findings on Kenya highlight the fact that identification with a certain entity, in the sense of a common identity, needs always to be understood in relation to a certain social and political context. We’s are a relational category in a double sense. First, in relation to the context, and second, in relation to other groups. Context matters; in everyday life the social scope of people’s future aspirations is the range of people include in their thinking and acting. This is a strong indicator of their particular ‘we’ or significant identity. Depending on the milieu, this ranges from ‘me’ to the nuclear family, the extended family, the church community, the ethnic group, or the nation. People whose future aspirations have a wide social scope combine different we’s in their action and thinking. When it comes to conduct of life, people relate to like-minded people. Of course, they do not talk about ‘milieus’, this is a sociological term, but they see similarities in their everyday life. When the social and political context changes, even those whose future aspirations have a narrow social scope will start to identify themselves with wider entities, such as their ethnic group. In Kenya, this is most obvious during times of in times of election with the ‘government camp’ or the ‘opposition camp’ during the post-election violence. The conflict on the regulation of abortion in the constitution created different camps, simplified conservatives versus liberal.

The second relational element is that each of the we’s also knows a ‘they’. In everyday situations, the social scope of people’s future aspiration does not include people for whom they do not feel responsible. Who is excluded depends again on the particular view of life and way of life—the milieu. However, in every case, there are they’s, those who are excluded. This becomes obvious in cases of conflict, in the moment of intransigence that is defined by a fundamental rupture of the whole of society. The relational constitution of we’s (and they’s) can be explained by referring to the insights of relational social science (Emirbayer 1997; Powell 2013; for African studies: Spies and Seesemann 2016). There are no fixed identities and we’s, because they vary according to the social setting and the wider social and political context. The changes of identification during the election violence in Kenya, from ethnic voting to two camps (government versus opposition), to the nation as an entity that should not be divided in violent conflict, shows that we’s are dynamic. Of course, the entities that constitute we’s and their social scope vary, and might be completely different in other countries. We’s can cross political borders, such as religious we’s or those of cosmopolitan liberalism, to name but two. However, the social constitution of views in a relational manner is a general feature. The identification of particular we’s is an empirical question that needs to be answered for every setting anew. We may learn from empirical findings from other settings and ask whether entities like class, ethnicity, nation, milieu, or smaller social we’s, such as nuclear or extended family or religious community, play a role, but we cannot take any of these entities for granted.