From Castes to Ethnic Group?
Modernisation and Forms of Social Identification among the Tharus of the Nepalese Tarai

Sigrun Eide Ødegaard

Institute and Museum of Anthropology, University of Oslo
Thesis submitted in partial fulfillment of the Cand. Polit. degree, 1997

Acknowledgements

This thesis is based upon fieldwork in the periods August 1993 - July 1994, November - December 1994 and January - April 1995. The Norwegian Research Council, the Nordic Institute of Asian Studies (NIAS) and the University of Oslo provided financial support. In the period 1993-1995, I was part of a project called "Civil Society, Environment and the Participation of NGOs in the Process of Democratisation in Nepal". The project was funded by the Norwegian Research Council and led by Harald O. Skar at the Norwegian Institute of Foreign Affairs (NUPI). Many thanks to NUPI for "letting me in" during this period, and for the inspiring "2 o´clock coffees". I would also like to thank NIAS for a two-weeks' library scholarship in March 1996.

In Nepal, I was affiliated with the Department of Anthropology and Sociology, at the Tribhuvan University, Kirtipur. I would like to raise special thanks to Dr. Ganesh Man Gurung at Tribhuvan University, for all kinds of assistance, both in the field and afterwards.

I would not have managed to accomplish this thesis without the help of my supervisors. My main supervisor, Arne Kalland (IMA), has guided me through the "trauma" of writing and helped me to structure my material. I also owe a lot to my secondary supervisor, Axel Strøm (IMA), whose knowledge of the South-Asian context (as well as much more) has been of great help and inspiration. Special thanks also to Eilert Struksnes (NUPI), who has "cleaned" my English and helped me with all kinds of formal, editorial "things".

In addition, I wish to thank the following people for their contributions:

Harald Skar (NUPI) for waking up my academic interest towards the Tharus and the Tarai
Inge Saghild for sharing with me her knowledge and experience of BASE, as well as for her great hospitality
Stein Tønnesson (NIAS) for inspiration and guidance
Ian Reader (NIAS) for fruitful comments to Chapter 6
Kari Anne Ulfsnes for fruitful comments to Chapter 6
Tove Cecilie Kittelsen for sharing with me "good times" and "bad times" in the field
Sarah Lund for guiding me through the initial stages of this project
Ingrid Sundt for introducing me to the "magic" of Power Point
Ole Dahl Gulliksen for spending hours (in vain) with my maps and figures
Lars Risan for technical and "print-out" assistance
Thomas Klevenberg for technical assistance
My parents for economic support at the very end
Geta Eyehospital and its staff members for hospitality and assistance during my times spent in the field
Dilli Bahadur Chaudhary and all of the BASE staff (in Tulsipur and in Geta) for sharing their time with me, and for answering all kinds of "strange" questions

Lastly, I wish to show my gratitude to all the people in Geti village, who patiently let me, "the white one" enter their houses and ask questions about their lives. Special thanks are raised to:

Yagya Raj Chaudhary
Chote Lal Rana
Gorya ghar, my Rana family for feeding and lodging me, and for caring for a strange daughter/sister/aunt.
Balmihana ghar, my Dangora home, for feeding me, and for their love, care and all kinds of assistance.
Sampatti Chaudhary, my "little sister" and assistant, who also became a good friend.

I would also like to thank Jeevan Rana and everybody in Surmeha ghar (Dhangadhi village), Gulaba and June Rana of the pradhaan's ghar (Rajpur village) for their assistance and hospitality.

"Blurred Boundaries"

"One can only interpret what one sees in terms of one's own experience and of what one is, and anthropologists, while they have a body of knowledge in common, differ in other respects as widely as other people in their backgrounds of experience and in themselves" (Evans-Pritchard 1951:84).

It was a Saturday afternoon in September 1993, and just a few days after my arrival in the far western district of Kailali, in the bazaar town of Dhangadhi. I was lying on my bed in the mothy room at the hotel I had booked into with my friend Tove. The monsoon had just ended, but it was still very hot, and because of the frequent electricity breakdowns, the fan did not work. Tove was sleeping, but I felt bored. What to do in Dhangadhi?, I wondered. I had already seen the last Hindi movie, so, despite the heat, I decided to go sightseeing on the bright new bike I had bought the previous day.

At the big crossing, Chauraha, a few kilometers north of the main bazaar, I decided to go to Kanchanpur, the neighbouring district. Instead of turning right, towards the East-West Highway, I turned left heading for the big river Mohana.

I soon came to something which looked like a national park. There was at least a big signboard which showed a tiger. I hesitated for a while, wondering whether I would have to pay some entrance fee of some kind. Nobody tried to stop me, however, so I continued my journey. The standard of the road was surprisingly good compared to other parts of the East-West Highway I had seen. I was also happy to see that rickshaws and buses passed regularly. This part was maybe not as remote as I had imagined, and while I cycled on, I was making plans for my work.

A voice screaming "madame, stop, madame, stop", brought me quickly back from my thoughts. I stopped and a rickshaw with two persons came up to me. They seemed to be very agitated. "Show me your passport", the more angry person asked me in stottering English.

I told him that my passport was in Dhangadhi, in my hotel room, and this made him even angrier, and he shouted "you wrong - you jail".

I realised immediately that I was dealing with the Indian border authorities, and that I actually was in India. Without an Indian visa, that is not the right place to be, and I became very worried. The two guys ordered me to come with them back to the border checkpoint. It is difficult to describe the desperation I felt there and then. I already foresaw my fieldwork being done in an Indian jail instead of in a Tharu village. A rather bleak vision.

On our way back to the border checkpoint, which I had failed to notice a few moments earlier, the rickshaw driver, with the heavy load of two grown-up men, moved much slower than I did. I therefore saw the possibility of reaching back to the border before my two "guards" and thus be saved. I started on a race back to Nepal, while my "guards" shouted behind me. I pedalled like crazy and had almost reached the border when a bus passed by, and unfortunately one of the men from the rickshaw jumped off the bus and grabbed my bike.

My hope was gone: This was the beginning of the deluge, I thought. In my mind I was already writing letters to my main funders, the Norwegian Research Council (NRF) and the Nordic Institute of Asian Studies (NIAS) explaining how the scholarships had been spent paying backsheesh to the Indian border police.

The guard took me by my arm, and brought me to a hardly visible "office" which was nothing but a table placed behind some trees. There was a very small signboard which said "border checkpoint". There was a lot of activity around this "office". Behind the table the chief officer was seated - the very stereotype of a North-Indian with his oiled hair and meticulously trimmed moustache. He seemed only too eager to start interrogating me. Why was I in India without a visa?

While my heart was beating, I explained as best I could that I had been on my way to Kanchanpur, and that I was a Norwegian student who had come to do research on the Tharu people in this part of Nepal. "How can you go to India without a visa?", the chief officer asked. I pointed out that I had no intentions of going to India, and that I had been on my way to Kanchanpur. He asked me again how I could be in India without a visa and after a second time, when I repeated my explanation, he asked: "don't you have a map?" As a matter of fact, I told him, while I tried to figure how much I should offer him to get out of this mess, I had not had time to get hold of a map. The chief officer, twinning his moustache, turned his head and looked at the audience of subordinates gathered around his table. To my surprise, he smiled and said: "She went to India by mistake."

This statement was followed by a collective nodding among all the subordinates. I was also quick to emphasise this. "Yes, yes, I went to India by mistake." "She went to India by mistake," he repeated and started giggling. This was followed by a collective giggling by his subordinates, and I smiled and nodded. Suddenly the chief officer said: "You should go back to Nepal." And of course I did not disagree with him in this. Before he let me go, he asked me to remember that he had saved me from being "tiger breakfast", and whenever I crossed the border in the aftermath of this incident, he would remind me and all others present about this.

This day, which initially had started with me feeling bored, had become much more eventful than I really liked. This remote part of Nepal thus seemed, to use Ardener's (1989) expression to be "full of events", and my fear of getting bored during fieldwork seemed to be "ungrounded". As a matter of fact, I never felt bored during my fieldwork, although this is not the reason why I have told this story.

The Tharu people of this part - "my people" - are bi-national. They live and marry across the political border. There is thus not only a geographical closeness between citizens in India and Nepal, but also a mental closeness. Tharus and Tarai people, it is often argued, are not nationally integrated. They are considered to be more India-oriented than Nepal-oriented.

The point I want to make, and which will be of interest in relation to the focus of my thesis, is the relationship between local, regional and national identities in the far western part of Nepal. With its geographical approximity, as well as its particular history, this part has had and still has a close relation to India. But to say that people here have no sense of national loyalty, is the same as to make an essentialist error. The boundaries between national and local/regional identities are context-dependent, Eriksen (1992) argues. The aim of my thesis is to describe and analyse the various forms and contexts where such identities are communicated. The boundaries between these contexts are, to quote Gupta (1995),"blurred".

I will not dwell any longer with this prologue, but rather start with the beginning; and introduce the problem and the main arguments I will forward over the next pages.

Introduction

Introductory Remarks

Throughout my thesis I will discuss what I see as two different, and in many ways, contrasting processes taking place among the Tharu population in Nepal. These are (i) caste-climbing and (ii) ethnic incorporation.

The ethnonym Tharu is attributed to several endogamous groups that are scattered all along the Nepalese lowland, the Tarai belt. The Tarai borders India in the south, and Tharus are also found in the adjacent areas of India. Until recently, the various Tharu groups were regarded as having little in common in a cultural and linguistic sense, and they looked upon each other as different kinds of people. These boundaries, however, seem to be in the process of disappearing, and many Tharus now argue that the Tharus are the same kind of people, and a growing ethnic incorporation is taking place. The pan-Tharu advocates reject their status as a caste (jaat) within the Nepalese caste structure, and instead emphasise that the Tharus are an ethnic group (samaaj) and an indigenous people (Adivaasi).

The pan-Tharu ideology is, however, not uncontested, and many Tharus still look upon themselves as castes and organise themselves according to the principles laid down in the Nepalese caste structure (see below). They distance themselves from the pan-Tharu movement and ideology and instead try to improve their social status by activities known as caste-climbing.

What I will discuss in my thesis, is whether the Tharus are in a process of changing from a group of castes into something like an ethnic group, and to what extent an ethnic incorporation is taking place. The process of ethnic identity formation is nothing naturally given, but has to be related to several processes, both external and internal ones. Therefore I find it useful to place the Tharus in a broader national and historical context, and relate them to Nepal's nationbuilding process which started in the mid-nineteenth century.

The Field: Tarai and its History

Nationbuilding and 100 Years of Rana Rule

After the unification of Nepal into a kingdom by the Gorkha King Prithvi Narayan Shah in 1769, Nepal went through a period of military expansion.⁽¹⁾ At this time, the different population groups of the kingdom were seen as both territorial and social units. They were both natives of a territory (des ) and members of a species/group (jaat ) (cf. Burghardt 1984).⁽²⁾

The Rana dynasty, a period of 100 years of hereditary prime ministership, deposed the Gorkha kings (Shah dynasty) from power in 1846. In 1854, Prime Minister Jang Bahadur Rana introduced the national legal code Muluki Ain which changed the way the government (and also people themselves) looked upon the various population groups. From that time, the various groups were solely referred to as jaat. This was both a product of, as well as a strategy for, Nepal's nationbuilding. Its two main purposes were to secure the country's independence from India and at the same time promote solidarity among the Nepalese citizens (cf. Høfer 1979; Burghardt 1984; Clarke 1995).

The main principle of the Muluki Ain was ascribed status, and it organised the nation's diversity according to a social model based on degrees of purity and impurity (Høfer 1979). It did not only include the native peoples of Nepal, but all within and outside Hindu humanity. Because of the Muslim and British take-over elsewhere on the subcontinent, the Nepalese government chose to pursue a national identity as the only Hindu kingdom of the world and the true Hindustan. This was articulated in a religious nationalism based on Brahmanic ideology (Van der Veer 1994. See also Kapferer 1988).⁽³⁾

In the Muluki Ain, people were classified into five groups (jaats), on the basis of the varna ideology⁽⁴⁾ known from Vedic Hinduism and the code of Manu.⁽⁵⁾ Three of these groups were considered pure and two impure (see figure 0.1 below). The pure groups were: (1) Wearers of the holy cord, tagadhari which were twice-born (dwij) of all nationalities, (2) non-enslavable alcohol-drinkers, namasinya matwali, tribals who could serve in the army, and (3) enslavable alcohol-drinkers, masinya matwali, tribals who could work as slaves in the houses of the twice-born).⁽⁶⁾ The impure groups consisting of various occupational castes were: (4) the impure but touchable groups, paani nacalnya, choi chito halnu naparnya, and (5) the impure and untouchable groups, paani nacalnya, choi chito halnu parnya. The ritually pure could not accept water from a person belonging to any of the impure jaats, but it was only people belonging to the lowest category who were considered "untouchables" (today commonly known as Dalits in India).

All of the three pure groups were incorporated in the Indian varna system, whereas the impure "water-unacceptable" groups were considered to be outside the four varnas. Most of the so-called tribals were grouped in the category Matwalis, either enslavable or not, and they were all ranked as shudra in the varna hierarchy.⁽⁷⁾ The social groups known as jaats, however, are groups based on kinship and marriage and not subsets of the varna categories. But although there is no automatic correspondence between varna and jati/jaat (cf. Quigley 1994), the hierarchical grading of the Muluki Ain was defined in terms of the varna ideology and based on the Brahmanic ideology of purity and impurity. This ideology has influenced most of the citizens, who, to a great extent, would make distinctions on the basis of - and think of themselves in terms of - the criteria laid down in this ideology (cf. Gellner 1989).

The Muluki Ain classified a complexity of social groups into substantial social categories, known as jaats. Every single group was represented as a separate caste, and subjective identity was ignored. The problem of what Höfer calls "internal status groups" was not clarified, and only among the so-called Pahaaris and the Newars were such internal groups described in detail (Höfer 1979). Due to its many meanings, the jaat concept is problematic, something I will come back to in more detail. Here, I will only mention that the term jaat is used as an ethnonym attributed to several different groups in the Tarai. The term is also applied to the various Tharu subgroups, such as the Rana, Dangora, Kochila and Kathariya. This was not specified in the Muluki Ain, where all these groups were subsumed under the collective term Tharu and represented as a single caste (jaat).

Although the Muluki Ain was abolished and deemed illegal in 1963, it still has relevance. The different jaats lived under this particular social model for such a long time that it has pervaded their own outlook as well (Sharma 1978). The Muluki Ain was based on orthodox Hindu values, and both religious power and political power were in the hands of the twice-born Pahaaris. Pahaari is a term commonly used on the Nepali-speaking hill population, who - despite linguistic similarity - is ethnically very heterogeneous. It consists of twice-born castes and occupational castes. The twice-born Pahaaris form the numerically, politically and culturally dominant part of the population.

Figure 0.1 The Caste Hierarchy of the Muluki Ain

 
Fall of the Rana Dynasty and Social Change in the Tarai

The Rana dynasty ended in 1950, and Nepal thereafter had a short period of parliamentary democracy. In 1960, however, King Mahendra declared the democratic system a failure. He dissolved the cabinet, arrested the ministers and in 1962 introduced the panchayat system, a "guided democracy" which was "more suitable to the Nepalese context". This system was composed of the king and four levels of councils (panchayat) elected at village, district, region and state levels. The national parliament (Panchayat Raaj) consisted of independent members, elected by adult suffrage as well as appointed representatives of different classes and professional organisations (Sever 1993). Both the nation-state and the national identity were based on the concept that all citizens were "one and the same" in a Vaishnavite devotional context.⁽⁸⁾ Political parties or any organisations on ethnic or religious basis were considered a threat to the nation's unity (Burghardt 1984).

After the end of the Rana dynasty, Nepal was opened up to foreigners, something which resulted in various development activities and rapid social change. The eradication of malaria in the Tarai led to a massive migration from the hill areas. Most of the new settlers were Pahaaris from the twice-born castes. From being an area mainly inhabited by Tharus and other malaria-resistant tribes, the Tarai became an ethnic cauldron. It has commonly been argued that the indigenous Tarai people have been exploited by these powerful newcomers, and that they feel powerless and marginalised in relation to their high-caste neighbours. Today, villages in the Tarai mirror this ethnic diversity, with Pahaaris and Tharus living in mixed villages. Geti village, where I did my fieldwork, is inhabited by three different Tharu groups (Rana, Dangora and Kunna) as well as the twice-born and occupational Pahaari castes. This setting gave me the opportunity to study inter-jaat relations on several levels.

Before I can go further, it is necessary to clarify and discuss some of the key concepts I will employ throughout this thesis. These are (i) caste, (ii) ethnic group and (iii) the concept "indigenous".

Conceptual Framework

What do we mean by Caste?

"We must search for that principle not in our minds, but in the minds of those people who practice the caste system, who have daily experience of it, and are thus most likely to have a feeling for what is most essential in it" (Hocart, quoted in Dumont and Pocock 1958).

It has been common to conceptualise castes as hierarchically ranked lineages which originally were associated with an occupation (cf. Kolenda 1978; Quigley 1994). As a form of social organisation, the caste system used to be linked to a local or regional system of interdependence and interchange (cf. the Indian jajmani system). There exists no single caste system, but local and regional variations of communities which arrange themselves hierarchically in relation to one another within a particular territory.

Caste - a Special Case of Social Stratification?

In a book edited by Leach (1960), several scholars raise questions about the character of the caste system. Is caste a type of social organisation peculiar to Hindu India, or are there structural elements in the caste organisation independent of Hindu cultural origins? To a certain point, Leach agrees that caste ought to be seen as a structural phenomenon. But he does not accept that it is a concept with worldwide application. To Leach, caste is indissolubly linked with pan-Indian civilisation (Leach 1960:5).

Barth, on the other hand, conceptualises the caste system as a special case of a stratified polyethnic system, which does not only serve to place individuals in discrete categories, but also provides for a hierarchical ordering of these categories into "higher" and "lower". If the caste concept shall be useful in sociological analysis, it must be based on structural criteria - not on particular features of the Hindu social order. Barth compares the system of social stratification among the people of Swat with the Hindu caste system. Although the people of Swat are Sunni Muslims, Barth considers their division into social groups known as qoum similar to castes. This similarity, he points out, is as a matter of structure rather than of culture (ibid.). Barth's view, which also is shared by Berreman (cf. Berreman 1979), makes it possible to find fundamental similarities in systems outside Hindu society.⁽⁹⁾

Other scholars have made a sharp distinction between castes and other forms of social stratification. In Dumont's famous Homo Hierarchicus (1980) - which has dominated and influenced the discussion of castes during the last decades - he disagrees with the general definition of castes suggested by Barth (1960). In Dumont's work, the caste system is closely related to the four varnas.⁽¹⁰⁾ Influenced by Bouglé's (1908) definition of the caste system, Dumont sees hierarchy, the opposition between pure and impure, as well as the dominant position of the priest, as embodying the real essence of caste and underlying the whole hierarchical organisation (cf. Dumont 1980). We can only talk about castes when all these features are present (ibid. See also Bouglé 1908).⁽¹¹⁾

Dumont does not dismiss the possibility of comparison, but he considers the caste system as an Indian institution with its full coherence and vitality in the Hindu environment. Some of the features that are constitutive of caste can, nevertheless, be present among other religious groups in the Indian environment. But the ideological fetaures mentioned above will - at certain points or in certain regions - be missing. To explain caste, therefore, is to explain why and when all of these ideological features are found together when only some of them are found elsewhere (Dumont 1980. See also Kolenda 1978; Quigley 1994).⁽¹²⁾

In this thesis, I will use Dumont's theoretical approach as a point of departure. Caste I therefore understand as an ideological system of thought where groups of people are ranked hierarchically according to degrees of ritual purity/impurity. This is expressed ideologically in rituals, but it also structures social relations (cf. Dumont 1980; Quigley 1994). There are thus restrictions on commensality between members of different castes, and in various contexts (such as those concerned with food, sex and ritual), a member of a "higher" caste may be ritually polluted by either direct or indirect contact with a member of a "lower" caste. A caste, (or a subsection of it), is, for reasons given above, usually endogamous. The caste system is, however, not a closed institution with no possibility of individual mobility (cf. Inden 1991; Srinivas 1996). Social mobility has been and is possible, and many studies have shown how various castes have improved their social status within the caste hierarchy. Such processes are known as "Sanskritisation" (cf. Srinivas 1996), which is explained as "the acceptance of the rites, beliefs, ideas, and values of the great tradition of Hinduism as embodied in the sacred books" (ibid.:76) and caste-climbing, where members of "lower" castes assimilate the customs and values of the higher twice-born castes (cf. Bailey 1970). Common for both "Sanskritisation" and caste-climbing, is that they are mostly in accordance with Vedic Hinduism and the classical varna ideology.

In a caste system, groups are ranked hierarchically according to their relative ritual purity. This ranking is defined on the basis of a common ideology, something which distinguishes a caste system from other multi-ethnic sytems. Although members of castes also sometimes describe their sense of distance from other castes in terms of culture or values, there is not much to gain in starting to call castes ethnic groups (cf. Searle-Chatterjee and Sharma 1994). "[T]o privilege 'cultural difference' over 'differential ritual status' would perhaps be to throw the baby of hierarchy out with the bath water of India's supposed uniqueness" (ibid.:19-20) and thereby deny the specific features of the caste system altogether.

Ethnicity and the Communication of Cultural Differences

The term ethnic group has become well-established in anthropological writings on Nepal. It has replaced the term "tribe", which was commonly used to denote culturally distinct groups, which were not integrated in a caste system. But, as Höfer (1979) points out, its substance has never been defined properly. Some authors on Nepal have used the term ethnic group as an antonym of caste or caste society in the same sense as the term "tribe" is employed with reference to India. (See below)

Like castes, membership in an ethnic group is also ascribed by birth. But as a form of social organisation, an ethnic group differs from that of a caste. Castes, I have argued, are ranked hierarchically according to ritual status, whereas ethnic groups regard themselves as culturally distinct social groupings (cf. Smith 1991; Eriksen 1992). An ethnic group is a form of social organisation, where membership is a function of "ascription and identification by the actors themselves" (Barth 1969a:10. See also Eriksen 1992), and where the members are conscious about forming a separate cultural and historical group (cf. Smith 1991). To Barth (ibid.), ascription by others is also crucial when it comes to ethnic membership. De Vos (1975), on the other hand, emphasises self-ascription as the most crucial aspect for any understanding of ethnicity.

The way members of an ethnic group mark boundaries between Us and Them, is by communicating cultural differences (cf. Eriksen 1992). Ethnic groups are thus primarily self-defining entities, and it is the communication of cultural differences which becomes crucial in the relation between ethnic groups, and not different ritual statuses. A transformation into an ethnic group may thus imply that distinctions which were formerly made with reference to a shared cultural framework of relative ritual purity/impurity, are abandoned in favour of cultural distinctions made on the basis of different cultural criteria. This is a central point which I will explore throughout the thesis.

Towards an Analytical Concept Empty of Substantial Content

The concept "ethnicity" has been used to explain such a great variety of social phenomena that Abner Cohen warned against the concept "becoming a fetish" (1982:307). "Ethnicity has already become the subject of such an extensive literature that there can hardly be any conceptual formulation about it not made by someone before" (ibid.:307). Since the concept in many ways became a "fetish" and used in popular discourse as well as in political rhetorics, some scholars have suggested to replace it with the more comprehensive concept of classification (cf. Eriksen 1992:6-7). Eriksen, however, defends the analytical use of the concept of ethnicity and suggests that it should be taken to mean "the systematic and enduring social reproduction of basic classificatory differences between categories of people who perceive each other as being culturally discrete" (ibid.:7). This is a formalist approach which goes back to the one suggested in Barth (1969b). "The critical focus of investigation is the ethnic boundary that defines the group, not the cultural stuff that it encloses" (Barth 1969a:15). This approach replaced the earlier substantivist focus on group characteristics and made it possible to view ethnicity comparatively. When I use the concept ethnicity analytically, it is in this meaning of the term. (It is, however, important to distinguish between an analytic approach to ethnicity and an emic perception, which is often substantivist and essentialising.) The ideas launched in Barth (1969b) led to many theoretical and methodological contributions on how one best can study ethnic phenomena.

When it comes to the analysis of ethnic identity formation, for instance, the subjective rationale for ethnic allegiances is emphasised. "The question is not, as Moerman points out, 'Who are the Lue?'... but rather when and how and why the identification Lue is preferred" (Moerman 1965:160). To the Lue, on the other hand, the question of who they are may be crucial, something which indicates the need to separate analytically between ethnicity as ethnic identity formation and ethnicity as a an aspect of group mobilisation (ethnic incorporation).

Some scholars consider ethnicity as a political phenomenon and relate ethnicity to group competition over scarce resources.(cf. Cohen 1974; Despres 1975a and c). "The assertion of ethnic status identities", according to Despres, "may provide an ideological basis for the corporate or political organisation of ethnic populations" (Despres 1975a:193). Roosens (1989) also focuses on the manipulation of ethnic identities for political purposes. Ethnic groups, according to Roosens, are "pressure groups with a noble face"(ibid.:14).

An emphasis on the political aspect of ethnicity is useful when it comes to the pan-Tharus who now make ethnic membership relevant for their political activities. Processes whereby ethnic factors become increasingly central for mobilisation and joint political action are commonly termed ethnic incorporation (cf. Eidheim 1971b; Handelman 1977). The political purposes inherent in the pan-Tharu movement is something I will come back to in Part Two.

To see ethnicity as solely a political phenomenon is an oversimplification.

Ethnicity is also tied to aspects of meaning (cf. De Vos and Romanucci-Ross 1975), which may vary according to contexts. Many scholars started to emphasise the social situation when studying ethnicity. Eidheim (1971b), for instance, analyses the different social situations in which ethnicity and ethnic phenomena occur among the Sami people of northern Norway. (See also Okamura [1981] for the relevance of a situational approach in studies of ethnicity and ethnic relations.)

In my analysis of Tharu ethnic identity formation, the question of self-ascription becomes fundamental. According to Guneratne, Tharu ethnicity exists on two levels (cf. Guneratne 1994). One is a modern pan-identity, and the other is locally tied to the particular endogamous group. This corresponds to what McDonaugh has described as a modern and traditional form of identity among the Dangora Tharu (cf. McDonaugh 1989). The new, modern, explicit form of identity is restricted to a small part of the population, mostly the young and educated. The traditional form, which is weaker and more implicit, prevails among the majority of the population (ibid.).

There are thus different levels of Tharu ethnicity, and the articulation of Tharu ethnic consciousness and ethnic commitment varies greatly between the modernised Tharu elite and the majority of the Tharus on a grassroot level (Guneratne 1994). In order to compare these various levels of social identification - an identification with a localised group versus an identification with a more abstract "imagined community", it is necessary to observe and compare how Tharus on a local level relate to, and express, their identity. By observing inter-ethnic and intra-caste encounters in Geti, my village of residence, I could grasp the various forms of articulation and thereby come to terms with the subjective meaning lying behind (cf. Eidheim 1971a; Berreman 1975; Eriksen 1992) . Forms and contexts for social interaction will be discussed in chapters 3 and 4.

Among the Tharus, as in many other cases (see Brass 1974), the elite are the principal agents in the process of ethnic incorporation. Mobilisation of ethnic groups is often governed by leaders with a political entreprise to forward and not necessarily an expression of the cultural ideology of the group or popular will (cf. Barth 1969a). It is therefore important to study the role of the entrepreneur in ethnopolitics (ibid. See also Thuen 1982). The Tharu elite possesses what Smith (1992) has called "ethnic consciousness", something which he explains as the stage in which a group knows about and manages to communicate shared myths and historical memories (ibid.). I will introduce the Tharu elite and its strategies in Chapter 5. But before I go any further, it is necessary to look closer into the Adivaasi concept, a concept which has become important for the Tharu elite in their political activity. Adivaasi is in the South Asian context used as a synonym for aboriginal/indigenous. The Tharu elite do not only represent themselves as an ethnic entity - they also define themselves as an indigenous group, Adivaasi, and identify with indigenous people elsewhere in the world. (The emphasis on a Tharu cultural distinctiveness does not prevent these actors from identifying with a much more abstract "imagined community" on a global level.)

Both caste and ethnicity can be studied as schema for exclusion and inclusion. Ethnicity is a kind of social identification where inclusion/exclusion are founded on distinctive cultural principles. These distinctions are considered to be based on different cultural criteria, which are complementary and not necessarily hierarchically ranked. Membership in an ethnic group is, to a great extent, based on subjective identification (cf. Barth 1969a). Castes, on the other hand, are groups which are hierarchically ranked according to their relative ritual purity. Although there might be disagreements about each group's status within a caste structure, these are nevertheless based upon a shared cultural and ideological framework. Moreover, membership in a caste is not so much based on subjective identification as membership in an ethnic group, but it may be a social category one is ascribed to by others.

Tribes and Indigenous Peoples

"After all, most people in India and Nepal were, if one could push history back far enough, tribal in origin" (N. Allen 1982:198).

Adivaasi, which literally means "first settler" (adi = first, vaasi = settler), is a term used to define what was earlier known as "tribes". In India and Nepal "tribes" were distinguished from "Hindus" (cf. Guneratne 1994; Bates 1995). The groups which consider themselves to be Adivaasi correspond thus to a large extent with those labelled "tribal".

A number of groups within the hierarchy of the Muluki Ain (the Matwalis) are usually denoted as ethnic groups in the anthropological literature on Nepal. These groups were earlier talked about as "tribes". Half of the population in Nepal was described as "tribal", whereas the other half was described as "caste" or "Hindu" (Guneratne 1994). There was, however, no traditional indigenous way of referring to the distinction between "caste", "Hindu" and "tribe", nor between "caste" and ethnic group.⁽¹³⁾ It was western research and administration that introduced the distinction between caste and ethnic group in South Asia (Höfer 1979; Bates 1995).

A common way for scholars to distinguish "tribes" from "castes", was in terms of social organisation. "Tribes" were associated with a particular territory and considered to "subsist in isolation" (cf. Bailey 1961; Sinha 1965; Kolenda 1978) in the sense that they were not integrated with others in a caste system. "Tribes", in other words, were conceptualised as territory-bound homogenous groups living in isolated villages (cf. Kolenda 1978). Furthermore, in contrast to the hierarchically ranked caste groups, "tribal" societies were considered to be egalitarian and undifferentiated (cf. Unnithan 1994). Another criterion which frequently reappeared, refers to the treatment of women (cf. Gellner 1989). Female roles and statuses are often central when it comes to the ways in which members across and within "castes" and "tribes" describe themselves and others (cf. Unnithan 1994).

Many scholars have ordered "caste" and "tribal" modes of organisation chronologically, and described in an evolutionary perspective the process whereby a "tribe" became a "caste" (cf. Kolenda 1978). The idea was that the "tribe" existed first, but as Hindu village life was established, the tribe became a caste. This evolutionary dichotomy has been abandoned, and it is now common to consider castes and tribes as categories of a continuum rather than separate dichotomies (Gellner 1989). Tribes, castes and ethnic groups coexist, and they even coexist within one and the same community, such as the Santal (cf. Orans 1965). The Hindu-tribal synthesis, Sharma (1978) points out, is a fact of Nepal's historicity.

The Tharu elite not only refuse their status as a caste, but they also claim to be an indigenous people and raise political claims as such vis-à-vis the government. When it comes to the term "indigenous peoples" as a self-chosen form of identity, it is largely a result of its currency in contemporary international legal and institutional activities (cf. Gray et al. 1995).

After the Indian independence, for instance, various systems of reservations for low castes, tribes and other "backward" castes and classes have been introduced. This is known as the Indian quota system. The Tharus in India were given special treatment according to the various classifications made by the authorities. In 1950, they were classified as a "scheduled caste", whereas in 1954 they were categorised as a "backward class". Since 1967, the Tharus - together with five other "tribes" in Uttar Pradesh - were declared an indigenous people and "scheduled tribe" (Adivaasi). The Tharus of Champaran (India) first refused to be labelled Adivaasi but later saw the advantage by being classified as such (Guneratne 1994). Many other so-called "tribes" have started to call themselves Adivaasi and mobilised politically into what is known as the Adivaasi movements. The term Adivaasi has more and more taken root and is used for tribe or ethnic group in India, whereas jati is generally used with the meaning caste (Höfer 1979).

"Indigenous" as a Political Category

The term "indigenous" is a political category (cf. Gray 1995), and we have to understand the term as "a political tool operating as an imperative term within a growing social movement" (ibid.:57-58). In order to prove their Adivaasi status, it has become important to the Tharu elite to trace a non-Hindu and thereby "tribal" past. A "new" myth of origin has therefore come in as an important argument, and the Tharu elite go back to old written sources where Tharus have been represented as a "primitive tribe" which was not yet "Hinduised". (I will discuss this in Chapter 6.) The eclectic nature of Hinduism, however, makes it difficult to distinguish Adivaasi communities from Hindu peasant communities in terms of their religion (cf. Bates 1995, see also Chapter 6). Many Adivaasis were once regarded as Hindus, and only because they found it advantageous, started to call themselves Adivaasis (Sinha 1962; Bates 1995).

Barth has pointed to the important role global discourses have come to play in processes of identity formation. When, for instance, indigenous people struggle in order to achieve a negotiating position, it is very often global discourses which define the arena (cf. Barth 1994). Ethnic revitalisation and ethnopolitical mobilisation are often results of globalisation and processes of modernisation, understood as the importance of capitalism, the overarching role of bureaucratic institutions and the growth of mass education (Eriksen 1992:134. See also Gellner 1983; Smith 1986; Anderson 1991). And changes in ethnic identity are often precipitated by radical changes in the political contexts in which people live (cf. Keyes 1982a:27). After the political and democratic shift in 1990 it has become possible and meaningful for Tharus and other communities in Nepal to link themselves to external global discourses, such as the ones on human rights and indigenous peoples' rights. A democratic system has also made it possible to form organisations based on ethnic principles. With its political and legal implications, the concept adivaasi has become useful for political actors. The term tribe, which denotes a culturally distinct and localised, territory-bound group, does not have the same political and legal implications.

As already pointed out several times, the emic term jaat covers both caste, sub-caste and tribe/ethnic group. The Tharus have been described according to the various understandings of the term. Jaat, in other words, is a problematic concept which needs clarification.

Jaat - Caste or Ethnic Group

According to Höfer, the Muluki Ain uses the term jaat with four different meanings:

1. It denotes taxonomically distinct groups with differing status positions within the hierarchy. It is only in this sense that jaat is rendered by caste (Höfer 1979). The Pahaari castes are best described in the Muluki Ain and individually mentioned. The Pahaari Brahmins are the highest ranked Brahmins in Nepal. The Thakuri castes, from which originates the royal dynasty, are ranked second, whereas the Chhetris, who dominate among higher officials and army officers, are ranked after the Brahmins and Thakuris (see figure 0-1).
2. Jaat can also mean a "caste-internal status group". Such a group does not necessarily correspond to a commensal or endogamous unit (jaat as defined in 1). The Tharus as a whole form one jaat, but the various endogamous Tharu groups (Rana, Dangora, Kathariya etc.) are also conceptualised as jaat. Even within these endogamous Tharu subgroups, there are also hierarchically ranked subcastes. The Rana Tharus of the far west, for instance, have internally ranked subgroups known as kuri that are talked about and conceptualised as jaat.
3. In some cases jaat denotes an ethnically (culturally and linguistically) distinct group which can embrace several castes, such as the Newars. The Muluki Ain speaks of the Newar jaat in the sense of ethnic group, but describes their internal stratification (also jaat) in the meaning caste.
4. Finally, jaat is used in the sense of legal status in general. (The Muluki Ain functioned as a legal code which regulated punishments according to jaat status.)

Guneratne suggests to replace the word jaat with the term ethnic group (Guneratne 1994:48). As they were laid down in the Muluki Ain, the jaat categories were not based on subjective identification. Although some jaats may be called ethnic groups, I find it problematic to start calling all of them ethnic groups. Subjective identification is crucial for membership in an ethnic group (cf. Barth 1969a; De Vos 1975; Eriksen 1992), and, according to Höfer, it is only right to use the term ethnic group on minorities who have a "subjective ethnic identity". That means to be "conscious of a solidarity due to a (mostly mythical) common ancestry and of sharing specific linguistic and cultural phenomena" (ibid.:47). Most commonly, this identity is expressed by an ethnonym, which often covers a certain local or regional range of dialectical and/or cultural features" (ibid.). The ethnonym Tharu, as it was used in the terminology of the Muluki Ain, represented the Tharus as a single caste, but this was not based on subjective identification, or dialectical or cultural features. Some of the effects of the Muluki Ain was, to quote Anderson (1991), that it made it possible to "say of anything that it was this, not that; it belonged here, not there" (ibid.:184). This taxonomy, which was imposed on top of the pre-existing local definitions, helped to create the current ethnic map of Nepal, where such identities were accorded importance (Höfer 1979; Clarke 1995).⁽¹⁴⁾ It functioned, to paraphrase Anderson (1991:184), as a "totalizing classificatory grid" which played a central role in reifying group identities in Nepal.⁽¹⁵⁾ Several examples from Nepal have shown the reinforcing effect the sanctioning of caste identities have had upon ethnic identity formation (cf. Höfer 1979:150). Although a collective pan-Tharu identity, one may argue, dates from the time of the Muluki Ain, it rests to be seen whether this identity becomes internalised as an ethnic identity most Tharus will identify with.

Whether the Tharus are transforming from various distinct castes into an ethnic group is the problem I will discuss in this thesis. To what extent is a process of ethnic incorporation taking place among the various Tharu groups in the Tarai? I will attempt to answer this by tracing the different ways in which identity is formed and expressed among the Tharus living in Kailali district and relate this to the formation of various ethno-political organisations.

I will now turn to the methodological approaches used in my study.

Methodological Approach

In the Field

Due to the large-scale immigration of Pahaaris (people from the adjacent hill districts), as well as by Tharus from the districts to the east, Kailali has become an "ethnic cauldron". I thought that this would give me the right setting for studying caste and/or ethnic relations among the Tharus, as well as their boundary marking vis-à-vis "Others".

In addition to the theoretical interests, there were also some practical considerations involved with my choice of fieldsite. The Nepalese antropologist Ganesh Man Gurung, (see acknowledgements) had earlier done fieldwork among the Tharus in Kailali, and he knew many people there. I had also been in Kailali during my previous visit in Nepal in 1991, where I had met "Madame",⁽¹⁶⁾ the Danish woman who for many years worked at the eye hospital near Dhangadhi, the administrative centre of Kailali district. This hospital was for long run by the Norwegian Church Aid. "Madame" has later also become much involved with a non-governmental organisation which runs a massive literacy campaign among the Tharus in the far west. This Tharu-based organisation is called Backward Society Education (BASE), and before I went to Nepal, I heard that one of the BASE offices was in Kailali. I was curious to see whether such a Tharu-based organisation would stimulate the development of a Tharu ethnic consciousness and thereby help in the formation of a collective ethnic identity (see e.g. Eidheim 1992). Would BASE be a kind of platform from which a collective Tharu identity could take form?

Arrival and Choice of Fieldsite

In August 1993, I arrived Kathmandu together with Tove Cecilie Kittelsen, a fellow anthroplogy-cum-Tharu student. It was in the middle of the rainy season, and due to the unhealthy climate as well as communication problems, we decided not to go down to the Tarai. Another and more important reason for staying in Kathmandu, however, was to acquire some basic knowledge of Nepali. Together with Tove, I had private lessons with a Nepali teacher every morning. And at Tribhuvan University, I found Tharu and other relevant literature which had not been available in Norway.

In late September Tove and I arrived at the airport outside Dhangadhi. Ganesh and some of his local friends were waiting for us, and, together with them, we visited several villages near Dhangadhi the next couple of days. To my disappointment, none of the villages we visited were "BASE villages", in the sense that they were running any of the BASE programmes. However, we came to know about a BASE office nearby the eye hospital. I told the BASE staff there that I was interested in their activities and that I would like to stay in a "BASE village". The head of this BASE office, who was very helpful, took me around to several such villages the following day. I finally decided to settle in Geti, the neighbouring village of the hospital. In this village, there were two BASE classes running in the evenings; one for women and one for boys. Inhabited by Dangoras, Ranas and Pahaaris, Geti mirrored the "ethnic cauldron" already mentioned. To me, therefore, Geti seemed to be just the village "I needed".

Before I moved into Geti, I had come to know several Dangora Tharus who worked either with BASE or at the hospital. In order to get better access to the Rana Tharus in Geti, I thought it was necessary to stay in a Rana house, and I moved into a Rana family of altogether 16 members. The houshold was known as Gorya ghar (Gorya house) and consisted of a senior couple, their two sons with their wives and children. Gorya ghar belonged to one of the better-off Rana families and was located near to three other Rana houses. The other neighbours were Brahmins, and one house was inhabited by Dangoras. I thought that it would be good for the focus of my study to live in an area where several jaats were settled.

The elected representative of the village, a Dangora Tharu by the name of Basu Dev Chaudhary, had insisted to negotiate about the rent and to help me in other matters. Basu Dev was teaching one of the BASE classes, and I came to know him through BASE. According to Basu Dev, I could not eat with this family, so I had to find another place to eat. He immediately offered me to eat in his home. Basu Dev is a religious devotee (bhagat).⁽¹⁷⁾ His family practices strict vegetarianism and are teetotalists. As the Tharus are known for their liberal food and drinking habits, I was afraid that too close a relationship with bhagats would have negative impact for my study.⁽¹⁸⁾ I also wanted to eat as "genuinely" Tharu as possible. Instead of eating in Basu Dev's house, I arranged to eat in the house of one of the Dangora girls working at the hospital. I thought it was a good starting point to have established my two "homes" among different Tharu groups, and that this would make my access more smooth and easy.

Living in a house does not automatically imply access and acceptance. And when it comes to the first period in my Rana home, this was not dominated by "inclusion". One reason for this was my lack of knowledge of the local Rana language. By the time I moved into Geti, I could to some extent manage without an interpreter and "speak directly with the people" in Nepali (cf. Srinivas 1996:229). But that did not help me much since many Tharus did not seem to understand my Nepali. None of the women in my Rana house, for instance, knew Nepali, a language which they referred to as pahaari bolii (Pahaari language). This was not so strange, since they all came from Rana villages in India. The old mother, whom I hereafter will refer to as aya (the Rana term for mother), was from a Rana village near the border. In the beginning, therefore, the women were talking about me - not to me - something which made me feel ill at ease. The two sons spoke Nepali fluently, but since they were busy in the fields, they were hardly at home.

The Rana language is similar to Hindi and thanks to the few Hindi classes I had taken at the University of Oslo, I could understand a little Rana. It was therefore not too difficult for me to learn some basic phrases. I could thus understand what people were talking about, and I was also able to say a few sentences myself. Although I never became fluent in the Rana language, my effort was important in the sense that I managed to "break the ice". I quickly experienced a change in the women's attitude towards me and felt more accepted.⁽¹⁹⁾ My communication with the Rana villagers was mainly in mixed Rana/Nepali. When the Ranas used words and expressions that I did not understand, there would usually be someone who found a Nepali equivalent, or a Nepali-speaking person who could translate. My main informants/translators, however, spoke Nepali fluently.

Participation and Observation, Watching and Listening

In participant observation, "the rule of method becomes the plastic, spontaneous faculty of application" (Skar 1991:12). It is commonly stated that a fieldworker is varyingly "observing" and "participating". My fieldwork was no exception from this rule - hence a mix of participation and observation, watching and listening. I did, for instance, not work in the fields. My Rana family told me that it was too heavy work for me, and they also thought that I might hurt myself on the scythe. And, I have to admit, that after I tried this hard work, I was quite happy to refrain from such activity.

It is only during the busiest agricultural periods that most people are in the fields, and even during these peak seasons somebody has to stay at home to cook and look after the youngest children. The women and the out-of-school children were usually at home, and the women would frequently sit down to talk and smoke. Sometimes I would help them with minor tasks, such as to carry oil seeds and vegetables up on the roof to dry, or to sweep the courtyard. During festivals, my role as participant was more dominant. At Dashain, an important festival celebrated by the Dangoras, I dressed like a Dangora girl and danced with the Dangora girls. The same was the case with the Ranas during their Holi celebration.

Thanks to my camera and tape recorder, I became a kind of village reporter, whose presence was wanted in weddings, funerals, ritual celebrations and village meetings. "Come on sister, bring your tape, there is a meeting", the villagers often told me. And if I had taped a meeting, the villagers wanted to listen to it at once. This was very useful for me, because I could ask again and get secondary/additional comments to - and observe how people reacted on - comments made during the meeting.

Contexts are emphasised as the crucial "thing" in ethnographic research, because contexts are so important for understanding "ongoing life" (cf. Hastrup 1989:7, my translation). To better understand "ongoing life", it is important to inscribe oneself into "otherness". Such an inscription is what deserves the term participation (Hastrup 1989:8). In order to grasp reality, it is therefore necessary to have experienced it, and herein, Hastrup emphasises, lies the concrete challenge of the ethnographer. The study of ethnicity implies a study of the social contexts for inclusion and exclusion. This was something I felt and experienced myself throughout my fieldwork. Although I do not base my analysis of Tharu ethnicity on my own experiences of being excluded/included, these experiences made me aware of the flux and fluidity in such processes.

Some of the first things I did after I had moved into Geti, was to visit each house in the village in order to get an idea of its size and "ethnic" composition. This was a good opprtunity to introduce myself to the villagers as well, and it was also important for my own feeling that I actually did something. I presented myself as a student who had come to learn about the Tharus and village life in general, and that I was interested in BASE and its activities. My presence in the BASE classes made me familiar with the curriculum used and the issues taken up for discussions. Through the classes, I came to know many of the women, something which made my access to their homes and families fairly smooth and easy.

I also have to say something about the time-span of my fieldwork. The thesis is based on three visits stretching over a period from August 1993 to April 1995. The most extensive fieldwork was done during an 11-month period, during which I mostly stayed in Geti village. My second visit was during the general elections in November/December 1994, where I also stayed mostly in Geti. These elections, however, gave me an opportunity to observe how "the local" was linked to larger, national issues. Thus, some of the ideas I had made after my first field visit had to be modified.

The third and last visit was made in the period from January to mid-April 1995. When I arrived Geti this last time, it was just after a big culture programme had been organised by BASE. Because of illness, I missed the programme itself, but the changes I noticed in the Tharu villagers, as well as the reactions and feelings they expressed, made me revise some earlier made conclusions. This also made me conscious about the importance "events" and matters from "outside" have on locals - even in so-called remote areas. During my three visits, I realised the limits inherent in a time- and space-bound study like mine. Societies are in constant flux, and my interpretation is the result of my interaction with "my" informants, which took place at a specific time and in particular contexts (cf. Geertz 1973). At the end it only rests to emphasise that all understanding is partial, and my understanding is no exception from this rule.

Multi-Levelled Linkages

In our attempt to understand and explain processes of ethnic identity formation, it is necessary to compare social interaction on a local level with national and global discourses and ideologies (cf. Barth 1994). I will here briefly say something about my approaches and techniques to the other levels I am discussing in this thesis: the pan-Tharu movement as well as the national indigenous movement. The empirical material from these levels does not come from long-term fieldwork, and it has therefore certain flaws and limitations.

The material which I use in my analysis is mainly based upon ideological formulations and public rhetorical statements by politically active people. During my stays in Kathmandu, I met and interviewed many of the actors who are involved in the national indigenous movement. These were both representatives of the ethnic umbrella organisation, the Janjaati Mahasangh, as well as representatives of other "ethnic organisations". These actors are internationally oriented, and they express their ideas in English-written newspapers and magazines. Some of them also write books and pamphlets in English. I was thus able to follow the discourse through secondary sources as well. But the majority of these activists I met during pan-Tharu conferences and meetings - arenas for the Tharu elite and the various Tharu organsiations. Next to the wide range of delegates I met and talked to there, I also had the opportunity to meet and talk to "ordinary people" who were not - to use Handler's expression - "professional ideologues"(cf. Handler 1988:32). Such a combination of remarks made by "professional ideologues" and those made by ordinary people therefore gave me an opportunity to grasp the complexities and variations in "ethnic commitment".

In addition to the process of "watching and listening", my material is also based on several informal interviews with government officials at various offices in Kailali district. (The Chief District Office (CDO), the District Forest Office (DFO) and the District Election Committee.) Next to the various British Gazetteers, Krauskopff's ethno-historical studies of the Tharus and the Tarai have been of vital importance to me. Although hardly anything has been written in English about the history of Kailali district, a Nepalese historian has recently written a paper placing Kailali in a larger historical context.

Visits at Lucknow University and Kumanyu University in Nainital made available most of what has been written about the Tharus in India, also more recent studies. In connection with the question about the Tharu origin, which has become an important issue to the elite, I visited several places in Rajasthan and was able to talk with historians at the City Palace in Udaipur, as well as at the University of Udaipur. I also made a short visit in the holy city of Hardwar, where Hindus all over the world (and also the Tharus) have their genealogies written down with the priests/scholars known as pandits. Last, but not least, a three-day Tharu Seminar which was arranged in June 1995, brought together old and young "Tharu scholars" from several countries. This seminar has been very illuminating for my own study.

Part One: Cultural Diversity

Chapter 1 The Tharus of the Nepalese Tarai

The focus of my thesis is on the Tharus, an indigenous⁽²⁰⁾ people of the Nepalese lowland Tarai. The "enigmatic Tharus" (cf. Lévi 1905) are considered to be the indigenous inhabitants of this former jungle area, an area where other people were reluctant to settle. According to the population census of 1991, there are 1.2 million Tharus, and they make up 6.5 per cent of the total Nepalese population of 18.5 million (Dahal 1994). As an ethnic category, the Tharus are among the largest in Nepal's multiethnic population and numerically ranked as fourth among the ten major population groups.

With a few exceptions, the great majority of the Nepalese people live in well-defined, specific geographic regions (Bista 1991). Tharus, who are found all over the Tarai belt, make up about ten per cent out of the total Tarai population of approximately ten million.

Map 1.1.Tarai Districts

The Tarai

The Nepalese Tarai, bordering West Bengal in the east and Uttar Pradesh in the west, is a vaste area more than 900 km long. The width varies between 50 and 90 km. The Tarai includes 20 of Nepal's 75 districts (see map 1.1.). The Nepalese Tarai is commonly divided into two zones, an inner zone known as bhitra madhes (inner Tarai) and an outer zone, the extension of the Indo-Gangetic plains, known as madhes (cf. Krauskopff 1989a; Guneratne 1994).

There is no geographical barrier between India and the outer zone. The Tharus are settled on both the Indian and the Nepalese side of the border, most of them in Nepal. The valleys of the inner Tarai are placed between the Siwalik hills (also known as Suriya) and the Mahabharata range, and they form a geographical transition between the plains and the hills.

The Tarai - the term used geographically on this part of the rain-forested plains - is a Hindi word meaning "feverish land" (Krauskopff 1989a). Sever (1993) traces the origin of the word Tarai to a Persian word, meaning "damp". Both etymological sources, derived either from Hindi "feverish" or Persian "damp", illustrate clearly the negative associations which for long were given this area. The term kalopaani (black/poisonous water) was also commonly used by other Nepalese, denoting its unhealthy climate, the malarious fever and the poisonous water.

But the Tarai has changed drastically over the last 30-40 years, both ecologically and with regard to its population. What until the 1960s was known to be a malarious rain forest area, with wild tigers, boars and snakes (Hamilton Buchanan 1819; Crooke 1896), has today become the breadbasket of Nepal. Among the main crops grown are rice, sugar cane, wheat, corn, pulses and mustard. Tobacco and jute are also grown, and there is a great variety of forest products.

Settlement and clearing of forest land in the Tarai started early in the 19th century (Cederroth 1995), but the Tarai remained a "remote area" (cf. Ardener 1989) in a national context, and a place most people were reluctant to move into. In order to increase tax revenues, which was the only income source for the government, the authoritarian Rana regime (1846-1951) decided to populate the Tarai area. First, they invited people from India to settle, and many Indian peasants from Bihar - for instance, the Yadav (one of the politically most important castes in Bihar) - settled and cleared the forest in the eastern parts of the Tarai (Ghimire 1992).⁽²¹⁾ The Rana government also liberalised the regulations related to crime, slavery and indebtedness. Criminals and runaway slaves were entitled to freedom if they settled and started to clear land in the Tarai. This settlement strategy, which was one of the first attempts to "nepalise" the Tarai region, was related to the government's concern to secure Nepal's independence from British India.

The early tax collectors in the Tarai, called chaudharys, were responsible for tax collecting in large administrative units, known as pargannas. The chaudharys, who were recruited from local elites, had a five-year contract from the government to collect taxes from a group of villages. A new tax-collecting system known as the jimidaari system⁽²²⁾ was introduced during Jang Bahadur Rana's rule in 1861 (cf. Cederroth 1995). The post of jimidaar was given to a chosen person, responsible for one village which had to give a fixed amount to the tax revenue office annually. After this fixed sum was paid, the surplus was the tax collector's own, and tax collectors would often become big and powerful landowners (jamindaars) (cf. Cederroth 1995). But because of malaria, very few landowners lived permanently in the Tarai, and a system of absentee landownership developed, where landowners came down to the Tarai only for a couple of months to collect the revenue from their tenants.

The Tarai was a popular hunting reserve for the ruling elite in Kathmandu, who often spent several months every year hunting there (Sever 1993; Guneratne 1994). During these hunts, local people had to assist as porters, elephant riders or in other labour-intensive activities.⁽²³⁾ After the fall of the Rana dynasty (1951), there was a period of social reform in Nepal. The country was opened up to foreigners, who were eager to start development projects in this poor country, which they nevertheless considered as a Shangri-La.⁽²⁴⁾ A malaria eradication programme started in the Tarai in the 1950s.⁽²⁵⁾

The jimidaar system was abolished with the land reform of 1964, a reform which allowed landowners in the Tarai to own not more than 25 bigha of agricultural land.⁽²⁶⁾ After malaria disappeared, the absent landowners moved down permanently. In order to politically integrate the Tarai people into the monarchy-led panchayat system, a government-implemented programme of resettling the Tarai with Nepali-speaking people started. The result was a "decade of destruction" of the Tarai (cf. Mishra 1990:14-15).

Large-scale migration from the neighbouring hill districts by hill people who were considered to be more loyal to the monarchy was encouraged. Ex-servicemen who had retired from the British-Indian and Nepalese armies, were especially encouraged to settle in the border areas. This, the government thought, would prevent the Tarai forests from being used as "sanctuary for arms raids and other political activities" (Ghimire 1992:65).

According to the population census in 1991, the Tarai had 30.9 per cent of Nepal's population.⁽²⁷⁾ The migration from the hills soon changed the Tarai into Nepal's breadbasket. With its subtropical climate and rich and fertile soil, the Tarai has 57 per cent of Nepal's cultivated land and provides 60 per cent of the country's total production. It is estimated that by the year 2001, 61 per cent of Nepal's population will be settled in the Tarai, and "this shift will transform Nepal from a classic mountain economy into a predominantly flat, subtropical and urban nation". (Himal Sept./Oct.1990:5-8). Cultivated fields and grassland dominate most of the former rainforest. The deforestation has been enormous and rapid. Whereas in 1927 about half of the total Tarai area was covered with forests, only one-fifth was forested in 1977. And today, there is hardly any forest left (Jha 1992, quoted in Gurung 1997). Most of the still existing forest is in the far west region, for instance in Kailali, which has 20 per cent of the country's forest (Gurung 1997).

The population increase in the Tarai is, therefore, not just a spontaneous response of the land-hungry peasants of Nepal, but a state-implemented process. Through a systematic resettlement of the Pahaari or hill people, the state attempted to culturally and ethnically transform this part of the country. This has been termed "Pahaarisation" (Shrestha 1990:167).

The Population in Tarai

The Tarai is today a mosaic of different population groups.⁽²⁸⁾ The rough division between Pahaari, Madhesi and Adivaasi is commonly used. Pahaari is a term used about the people traditionally settled in the hills (pahaar). The hilly nature is also commonly used as a metaphor for "the real Nepal". Madhesi means simply a person living in the lowlands, madhes, but refers to people of Indian origin. Madhesi has today negative connotations, and the term is very often associated with illegal immigrants from India. There is a strong feeling among Madhesis that they are treated as second-class citizens. The Madhesis are composed of various groups. There are Muslims and high-caste Hindus, as well as the various occupational caste groups. According to Dahal (1994), there are more than 30 different jaats among the Madhesi Hindus. The Muslims as well as the occupational castes are settled all over the Tarai belt (ibid.).

The Adivaasi (indigenous) groups of the Tarai are settled in various geographical areas. The Dhimals, Meche, Rajbhansi and Sattar live in the far eastern districts of Jhapa and Morang. Like the Muslims and the occupational caste groups, Tharus are settled all over the Tarai belt and they form the numerically dominant Adivaasi group in the Tarai (cf. Dahal 1994; Harka Gurung 1996).

The major languages spoken in the Tarai are Maithili (2.2 million), Bhojpuri (1.4 million) Tharu (1.0 million) and Awadhi (0.4). Maithili and its various dialects are dominant in the far east. Bhojpuri prevails in the central Tarai, whereas Awadhi is restricted to the far west. Various Tharu languages are spoken all over the Tarai region, but they are much influenced by the three languages mentioned above. Other languages spoken are Urdu in the west, Hindi in the central and Bengali in the east. They outnumber most of the Tarai "tribal" languages (cf. Gurung 1996).

Table 1.1. The non-Pahaari Population of the Tarai

Source: Harka Gurung 1996.

When dealing with the Tarai people, it is important to keep in mind the migration factor (Krauskopff 1989a). Because of wars, natural catastrophes and sometimes by order of the local kings, the population in the Tarai has been constituted by incessable migrations in search of new land. The Tharus also have a past characterised by migration, enforced by the deforestation which took place during the last 100 years. Already in 1885, Nesfield pointed out the consequences the rapid deforestation had for the Tharus (in India): "Since the advent of British rule, the forest has been disappearing with surprising rapidity, and the Tharus have retired closer than ever to the Naipal mountain" (Nesfield 1885:36). And Nesfield argues that the migration of the Tharus from Gonda district to Nepal started about a century earlier, i.e., late 18th century (ibid.:35).⁽²⁹⁾

As a whole, the Tharus from east to west are composed of various endogamous groups with different cultural practices and languages. In the classificatory system laid down in the Muluki Ain of 1854, these various groups were subsumed under the general term Tharu and placed among the lowest of the pure jaats (cf. Introduction).

Traditionally, the various Tharu groups looked upon each other as different from each other, and the different Tharu languages are often mutually unintelligible (Bista 1967; Guneratne 1994). There is, as we will shortly see, also a great variation in the way the various Tharu groups have integrated into mainstream Nepalese Hindu society. This cultural and linguistic variety, as well as the various Tharu groups' tendency to mutually exclude each other, makes it problematic to talk about the Tharu people as one ethnic group or ethnie (Krauskopff 1989a).⁽³⁰⁾ (See also Introduction.)

I will now take the readers on a journey from Mechi to Mahakali, through the land of the Tharus. The main purpose for doing so is to place the various groups in their geographical and historical context, as well as to give the readers an idea of the variety existing within the people known as Tharu. The East-West Highway (Purba-Paschim Rajmarg) links the various parts of the Tarai together, and I will start the journey by the border of Mechi river in the far eastern district of Jhapa. I will follow the much-used division of the Tarai belt into three parts: East, Central and West. Almost half of the Tharu population (47.6 per cent) are found in the western part of the Tarai. Approximately one third (35.2 per cent) is settled in the eastern and the far eastern Tarai districts, whereas 17.3 per cent of the total Tharu population lives in the central Tarai districts (Gurung 1996).

From Mechi to Mahakali: Tharu Cultural Diversity

"...the objects of study are not encompassing logical orders or structural wholes or internally homogeneous and shared, and externally bounded cultures but rather variations themselves and the processes and mechanisms that produce them" (Vayda 1994:323 ).

Entering the Nepalese border town Kakarbhitta from the Indian state of West Bengal, brings the visitor to one of the most densely populated areas of the Tarai. According to Dahal (1994), the Tharus in this eastern part are the numerically dominant group in Sunsari and Bara districts only, whereas Muslims and Yadavs, as well as different Pahaari groups dominate the other eastern Tarai districts. In the districts of Sarlahi, Mahotari and Dhanusa, there are only 20,590 Tharus. Further east, in Siraha and Saptari, the Tharu population is 82,257, whereas in Sunsari and Morang, the districts east of Kosi river, the number of Tharus is 145,070 (Gurung 1996).

East Tarai

The east has been politically and economically the most important part of the region and also the country's most important area next to Kathmandu (Sever 1993; Guneratne 1994). Already 150 years ago, Prime Minister Jang Bahadur Rana made efforts to increase his tax revenue from the eastern part of the Tarai, which was the most productive part of the region. Large yields, for instance, came from the districts of Morang, Saptari and Mahottari (Stiller quoted in Deuel and Meyer 1997). In 1809, the rent for pastures in Morang district gave the Nepalese government a revenue of Rs. 24,000, and the duty on Morang's timber was Kathmandu's second largest source of revenue (cf. Singh 1990:16).

Most social and structural changes in the Tarai started in the east. When, for instance, the Rana rulers decided to have a regular police force in the Tarai in 1914, they started with Birgunj district (now Bara). In 1927, a telephone line was established between Kathmandu and Birgunj (Sever 1993), and Nepal's only short railway line was also in this part. This stretch connected Nepal with the Indian railway network, and was important for the trade between the two countries. The Biratnagar jute mill factory was responsible for Nepal's main export to India.

Situated near West Bengal and Calcutta, the eastern Tarai was also important in a political sense. Supporters of the Nepali National Congress (a pro-democratic movement much influenced by its Indian counterpart) were detained in Biratnagar, Birgunj and Janakpur in 1947. The Koirala family lived near Biratnagar and was a driving force in the formation of the Nepali National Congress. Later, two of Nepal's prime ministers came from this Koirala family. The industry in the Tarai experienced instances of labour unrest as well, and 10,000 workers were striking at the Biratnagar Jute Mill in 1947.

Apart from Tharus and other native Tarai peoples, many Indian immigrants have settled here the last 150 years. The great famine in the adjacent district of Bihar in 1769-70, followed by other natural calamities, led to a huge immigration from the Indian side (Guneratne1994; Deuel and Meyer 1997).

According to the census of 1971, the Tharu population is absent in some of the eastern districts. The classification in Nepal's censuses is based on languages, and after Nepali, Maithili is the language most frequently used in the central and eastern Tarai regions (Dahal 1994). Due to intense settlement of Bhojpuri- and Maithili-speaking groups of Indian origin, 80 to 90 per cent of the people in these eastern districts regard Maithili and Bhojpuri as their mother tongues. The large population of Tharus in the eastern Tarai is often forgotten.⁽³¹⁾ People who earlier had been classified as Tharu, were - because of linguistic assimilation - classified as Maithili/Bhojpuri. It was the image of the Tharu in the far west that came to represent Tharus as a whole (Guneratne1994).

Unlike many other jaats in Nepal, there are not many Tharus who actually use the category Tharu when they refer to themselves. Bista (1967) mentions several different names used by the Tharus in the east, who today are collectively known as Kochilas.⁽³²⁾ Many Tharus in the east are also called and call themselves Chaudhary. Chaudhary was a title originally given the land revenue collector in the Tarai. The high status accorded the Chaudhary title resulted in a process of "Chaudharisation" among Tharus all over the Tarai.⁽³³⁾

The Tharus' close contact with high-caste Hindus was followed by an emulation of high-caste values (Bista 1967). Bista described "a wave of reform" taking place among educated young Tharus, who "changed their food habits, reformed their religious practices and introduced modern education" (ibid.:127). Bista also mentioned the Tharu Welfare Society - an organisation which encouraged education among the Tharus and provided hostels in Birgunj for school children and students of both sexes. (See Chapter 5.)

As a result of this long tradition of modernisation, many Tharus in the east are today working as schoolteachers, in public administration or in other white-collar jobs. They do not till their own land, and, often, the low-caste group of Musahaars (literally rat eaters), who are migrants from India, work as day labourers on many of the Tharu landowners' fields. These Musahaars, according to Deuel and Meyer (1997), are in many cases indebted to their Tharu landlords. This, they point out, is the reverse of the situation in the western part of the Tarai, where Tharus are often found as bonded labourers (kamaiyas) to high-caste landlords.

The eastern Tarai has many places of religious and mythical importance. Janakpur in Dhanusha is the birthplace of Sita, the faithful wife of Rama and the main protagonist in the great Hindu epos, the Ramayana (Valmiki, 300 BC). The temples of Rama and Janaki in Janakpur are holy places commonly visited by pilgrims. The Tharus in the east have been integrated into Nepalese mainstream society for quite a long time. They speak Nepali, dress in the same type of clothes which are common in most of Nepal, and live in multi-caste villages where they participate in and share the same rituals as elsewhere in mainstream Nepal (Guneratne 1994:58).

Central Tarai

The district of Chitwan in the western part of the Narayani zone, together with the districts of Navalparasi, Rupandehi and Kapilvastu in the Lumbini zone, make up the central part of the Tarai. A common name for the Tharus of Chitwan and Navalparasi is, according to Guneratne (1994), Chitwaniya.

The Chitwan valley is the largest of the broad valleys north of the Siwalik hills (Müller-Böker 1991) and is today well known for the the Royal Chitwan National Park, which, next to Kathmandu and mountain trekking, has become a place for tourists in Nepal. The National Park - established in 1973 - covers more than 1,000 km2 of the forests and grasslands. These natural resources have been withdrawn from use by the local population, and the Tharus are those who have been mainly affected by the loss of these forests and grasslands (cf. Müller-Böker 1993).

In 1939, Chitwan was characterised as "the fever hell of Nepal", where "human beings are living, but they are starving apparently between life and death" (cf. Müller-Böker 1993). During the last decades, a total transformation has taken place in Chitwan, which has developed from a "sparsely populated periphery to an attractive multiethnical centre" (Müller-Böker 1993. See also Guneratne 1994). When the official programme of "Pahaarisation" started,⁽³⁴⁾ the first resettlement was established in the Chitwan Rapti valley in 1953. At about the same time, development programmes were also implemented, and, according to Müller-Böker (1993), Chitwan has been free of malaria since 1969.⁽³⁵⁾ At present, the inhabitants of Chitwan are mainly Pahaaris who bought land in order to profit from the tourist business. The busy bazaar of Narayanghat, which is the highway crossing and main centre in Chitwan, is today an example of how Chitwan, to paraphrase Müller Böker (1993), has changed from a "fever hell" to a "melting pot". 70 per cent of the Chitwan population are Nepali-speakers, and Tharu as a mother tongue is only spoken by 12 per cent (ibid.).

There are many examples of how this resettlement system was exploited by landowning Pahaaris in order to acquire large land tracts in the Tarai.⁽³⁶⁾ The rich soil and high productivity of the land encouraged many wealthy landowning Pahaaris to buy land from Tharus. The construction of an irrigation canal in the 1960s, as well as roads which made access to markets easier (Ghimire quoted by Skar 1992) gave further strength to Pahaari exploitation. Some Pahaaris also encroached illegally upon Tharu-owned land, which they later were able to register in their own names. The Tharu farmers thus became losers in a system which was implemented by the Government. A result of this is therefore a conflict between the indigenous Tharus and the estate-owning Pahaaris (cf. Skar 1992; Müller-Böker 1993).

In Kapilvastu district, west of Rupandehi, lies Butwal, which is the third biggest town in Nepal and also an important trade centre. According to Guneratne (1994), the Chitwaniya Tharus are related to the Tharus in the Butwal area, Tharus who are known as Kathariya. Kathariya Tharus are settled in the area from the central Tarai and Gorakhpur in India all the way westwards to Kailali (Krauskopff 1989a). The Kathariya Tharus are divided into two mutually exclusive groups; "the western ones", called Pachal and "the eastern ones", called Purbya (Krauskopff 1989a).

West Tarai

The western part of the Tarai consists of four zones - Rapti, Bheri, Seti and Karnali - and Nepal ends in the west by the border town Banbasa near the Mahakali river.

The neighbouring district west of Kapilvastu is Dang-Deukheri, the place of origin of the largest Tharu group in the west, the Dangora Tharu. Dangora is a term derived from Bhojpuri and means "Tharu from Dang" (Krauskopff 1989a). The East-West Highway crosses through the more fertile valley of Deukheri through which the river Rapti flows. The Deukheri valley is situated at an altitude of about 300 m, and was, according to Krauskopff (1989a), populated later than Dang. The Tharus of Dang-Deukheri consider themselves to be of the same group. Krauskopff notes, however, that there are several differences between the Tharus of the two valleys, especially when it comes to social and economic issues. Tharus in Deukheri, for instance, tend to be richer and more powerful in their relation with the Pahaaris. And the relationship between inherited priests and clients which is so fundamental in Dang, does not exist in Deukheri. These differences, Krauskopff notes, do not prevent these Tharus from considering themselves as one group.

The Dangoras are originally from the valleys of Dang and Deukheri, but today we find them in all the five far western districts, as well as in the Gonda and Bahraich districts of India (Krauskopff 1989a; Guneratne 1994). Due to immigration from the neighbouring districts, deforestation and natural calamities, Tharus in Dang-Deukheri have faced severe problems, the most severe probably being bonded labour. An organisation called Backwards Society Education (BASE) has been formed among Tharus in Dang to fight against bonded labour, which is commonly known as kamaiya.⁽³⁷⁾ According to a survey on bonded labour, which was