Ødegaard, Sigrun: From Castes to Ethnic Group? Moernisation and Forms of Social Identification among the Tharus of the Nepalese Tarai

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

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".

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:

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"

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

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.

1. Caste group of the "Wearers of the holy cord" (tagadhari)

Upadhyaya Brahmin
Rajput (Thakuri) ("warrior")
Jaisi Brahmin
Chhetri ("warrior")
Newar Brahmin
Indian Brahmin
Ascetic sects
Various Newar castes

2. Caste group of the "Non-enslavable Alcohol-drinkers" (namasinya matwali)

Magar
Gurung
Sunuwar
Some other Newar castes

3. Caste group of the "Enslaveable Alcohol-drinkers" (masinya matwali)

Bhote (people of Tibetan cultural extraction)
Chepang
Kumal (potters)
Tharu
Gharti (descendants of freed slaves)

4. Impure, but "touchable" castes (Paani nacalnya choi chito halnu naparnya)

Kasai (Newar butchers)
Kusle (Newar musicians)
Hindu Dhobi (Newar washermen)
Muslim
Mlecch (Europeans)

5. Untouchable castes (Paani nacalnya choi chito halnuparnya)

Kami/Lohaar (blacksmiths) and Sarki/Chamaar (tanners/leatherworkers)
Kadara (stemming from unions between Kami and Sarki)
Damai (tailors and musicians)
Gaineb (minstrels)
Badi (musicians and prostitutes)
Cyame (Newar scavengers)

Source: Guneratne (1994: 31), modified after Höfer (1979:45). The Muluki Ain uses the single term jaat to refer to the various groups that Western social scientists describe by terms such as caste, ethnic group or tribe (Høfer 1979:46).

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.

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.

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.

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).

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:

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.

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?

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.

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.

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.

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 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).

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).

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).

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).

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 done by BASE (cf. Cox 1994), every bonded labourer in the far west is Tharu. And of the landlords, 97 per cent were twice-born Pahaaris (Brahmins, Chhetris or Thakuris).

A massive outmigration of Dangoras from Dang-Deukheri and westwards has taken place. The last 30-40 years, 6,000 Dangoras have migrated westwards from Dang valley (Gunreratne 1994). From having formed a vast majority in Dang, the Dangora Tharus suddenly comprised only about 45 per cent of the total population (cf. McDonaugh 1984a). Due to this migration, the increase in the Tharu population in the western districts of Kailali, Bardiya and Banke has been more than 22 per cent. The Dangoras are the largest Tharu group in the far west, and make up approximately 70 per cent of the total Tharu population (Krauskopff 1989a). According to Harka Gurung (1996), most of the Tharus in the far west are settled in Kailali district. 206,933 of the people counted were Tharus, a number which makes up 49.5 per cent. In Bardiya, the 153,322 Tharus make up 52.8 per cent of the inhabitants counted, whereas in Kanchanpur, the 70,544 Tharus only form 27.4 per cent.

Many of the Dangoras who are settled in Bardiya and Banke, have migrated from Dang-Deukheri. Their linguistic and cultural practices resemble the Dangoras in Dang (Krauskopff 1994, private conversation), even though the Dessaria Tharus I talked to in Bardiya claimed that they were a different jaat. The main Tharu groups in the far west mentioned so far, seem to be related to the Dangoras further east.

There are, however, other Tharu groups in the west. The Kathariya Tharus already mentioned are also settled in Banke, Bardiya and Kailali (Krauskopff 1989a; Deuel and Meyer 1997). The Tikapur area in the eastern part of Kailali is inhabited by Dangora Tharus and Kathariya Tharus. According to Krauskopff (1989a) these Kathariyas are different from the Kathariyas in the Butwal area, and have several common features with the Rana Tharus of west Kailali and Kanchanpur, such as the female dress and the house style. The name Kathariya is probably derived from their occupation of bringing wood (kaath) on the rivers down to India (Deuel and Meyer 1997).

Bardiya district ends by the Karnali river, where Kailali district starts. West Kailali and the territory all the way to Nainital in Uttar Pradesh are known as the Rana Tharu area. Ranas are today settled in Kailali and Kanchanpur districts of Nepal, as well as in Kheri and Nainital districts of Uttar Pradesh. For a long time, the Rana Tharus were the best known Tharu group in an ethnographic context. Rana Tharus in India have been mentioned in literature (e.g., British Gazetteers) from the last century, and in 1958, the Indian anthropologist Srivastava published his monograph on the Rana Tharus of Nainital. The Rana Tharus came in many ways to represent the Tharu as a whole.

The former malaria-infested Tarai has in short time changed into the breadbasket of Nepal. From originally being inhabited by malaria-resistant "tribes", such as the Tharus, the Tarai is today an ethnic cauldron. The population increase has been tremendous. This has not been just a spontaneous response of the land-hungry peasants of Nepal, but a systematic state-implemented process commonly known as Pahaarisation (cf. Shrestha 1990).

In this chapter I have described how the changes affected the Tharus all over the Tarai. Tharus form the numerically dominant Adivaasi group in this part of the country. The Tharus as a whole consist of various endogamous groups with little in common in a cultural and linguistic sense. In order to better understand this cultural diversity within the Tharu population, I made a travel over the Tarai belt and placed the various groups in their particular historical and geographical context. The various parts of the Tarai have played different roles in a national sense. The east, for instance, was an important part of the region, both politically and economically, and the Tharus in this part, have for long been influenced by Nepalese mainstream culture.

Since the eradication of malaria started in the 1950s, as well as the establishment of the Royal Chitwan National Park, the population pattern in the central part of the Tarai has totally changed. Various development programmes, next to the tourist business, encouraged a lot of Pahaaris to settle there.

Because of problems faced in the valleys of Dang and Deukheri, the Dangora Tharus migrated westwards. This migration has led to an increase in the Tharu population of the far western districts. The Dangoras today outnumber the original Tharu inhabitants of the far west - the Rana Tharus. I will in the next chapter introduce you to the far western district of Kailali and take a closer look at the Tharus of the "Far West".

Chapter 2 Kailali District - the Land and the People

A Remote Area

Until the creation of modern Nepal, most of the Tarai was divided between its neighbouring hill principalities (cf. Krauskopff 1989a; Subedi 1995), and revenues had to be paid to the different hill kings (raajas). Kailali was under control of the various kingdoms in the Karnali region, and the hill kingdom of Doti administered and collected revenues here in a period before unification. Because of the trade which took place between the hill kingdom and the plains, there were several important centres in Kailali (Subedi 1995).

The far west never had the same importance for Nepal as the eastern part of the Tarai.⁽³⁸⁾ Most Nepalese citizens have not been west of Nepalgunj in Banke district, and this far western part of the country is considered a very remote area - an area which would fit in with Ardener's archetype of "remoteness" (cf. Ardener 1989:213).

While travelling through Kailali district, a feeling of "remoteness" is certainly still salient. Even though a big new bridge has recently replaced the former tow-ferry, the Kailali part of the East-West Highway goes through jungles and a rocky landscape.⁽³⁹⁾ In addition to a lack of modern infrastructure, the climate has also contributed to this feeling of "remoteness". During the rainy season, there is no link between this part and the rest of Nepal. Communication east-west is very difficult, and in order to go to Kathmandu, people have to travel through India.

The population structure here, like elsewhere in the Tarai, has changed drastically in recent years. Krauskopff (1989a) pointed to the fluctuation which characterised the Tarai people and their settlement pattern. Many stories I heard confirm this fluctuation and instability. Sometimes the population pattern of a village changed totally because of out- and in-migration. I will just mention two examples which illustrate how the settlement of two villages changed in a few years' time. East of Dhangadhi, the administrative centre of Kailali district, there is a village called Chaumala, which a few years ago was a Rana village of 117 houses. Today there are no Rana houses left, and Chaumala is today inhabited mostly by Pahaaris and a couple of Dangoras. The same is the case with Malakheti, a village west of Dhangadhi. A former Rana settlement is today inhabited by Pahaaris.⁽⁴⁰⁾ Many Ranas have moved to India, the Chief District Officer (CDO) in Dhangadhi told me. (According to this CDO, the Ranas went to India because of the special treatment they were given there as a "scheduled tribe/caste" in the Indian quota system See Introduction.)

This former jungle area has seen an enormous deforestation, which, according to Ghimire (1992), has been mainly politically motivated. Until 1925, Kailali had abundant forests, and in order to get land cleared, the government started to pay people 10,000 rupees pr. unit land cleared. If more than one unit was cleared, a five years' tax exemption was given (Gurung 1997).

What is at present the big road leading from the border checkpoint Mohana near Dhangadhi, was a couple of decades ago covered by forest and renowned for all the tigers. Another massive deforestation took place in the years between 1975 and 1981. Lots of people were engaged in the timber export to India, an export which was deemed illegal in 1981. In the period before the referendum in 1980,⁽⁴¹⁾ political leaders of the panchayat system were allowed to export timber so that they could finance their election campaign. A heavy deforestation also took place during the construction of the East-West Highway in 1985-86.

All the forest in Kailali is state-owned, and nobody is today allowed to cut trees for their own use. If a person is caught in illecal cutting, he is either fined or jailed, and there are 74 armed forest guards, known as bal bhutti, controlling the forests (Gurung 1997). These bal bhuttis are known for taking bribes. A replanting project has recently been implemented, and - according to the DFO in Dhangadhi - about 1,500 to 2,000 hectares of trees have been planted. There is a district-based research committee which makes decisions in relation to the planting of new trees.

As mentioned in the previous chapter, the Tarai was a popular hunting ground for the ruling elite, a place where they could hunt for big game, such as rhino, tiger, deer, sloth bear and elephant. Due to the deforestation, some of these species lost their habitat - hence the government saw the need to establish reserves and national parks. In the Nepalese part of the Tarai, there are today three national parks: in Chitwan, Bardiya and Kanchanpur. Across the border, in Kheri district in India, there is a national park called Dudwa.

The Tarai nevertheless still has an abundant fauna, but according to Krauskopff (1897c; 1989a) hunting was of almost no importance to the Dangora Tharu economy, an economy which was primarily based on farming and fishing. Some Tharus would put up traps for rats and various birds, but otherwise hunting was not a main activity. Srivastava (1958), on the other hand, described the Rana Tharus in Nainital as great hunters, and he listed the many and various hunting techniques and utensils used by them. "As the Tharus are great hunters, they, in spite of the forest laws, steal into the forests for their hunt" (ibid.:53). In the far west, I heard rumours of people illegally hunting species such as jungle fowl and jungle cow (nil gai). There is often big money involved in these illegal hunts. Tharus in my village of residence, however, told me that hunting was not very common nowadays. I never saw any adult men go hunting, but children would set up traps for birds, and many young and old would set up traps for the big rats (musa) usually found in the fields.

Fishing, on the other hand, is an important and a popular activity, and the fishing nets and techniques among the Tharus are many. The whole year, people can go fishing in the rivers, and during and right after the monsoon, people even catch small fish in their fields. Big nets are cast into the larger rivers, whereas smaller lift nets are used for fishing in irrigation canals and small streams. Other nets and techniques are also common.

The forest is often used for collecting berries, fruits and honey. During the season, people can collect wild papaya and mangoes from the trees which grow in huge quantities. Medicinal plants are also collected, and Dangoras use a particular leaf for making eating-plates and cups. Grass for house construction as well as reeds for mat-, rope- and basket-making are also collected. The soil is generally fertile, and now that irrigation has become more common, water does not seem to be a big problem.

The population of Kailali district has increased dramatically, and the annual population growth in some areas of Kailali is eight per cent (cf. Skar 1992). According to "Kailali's king" (a big landowner who was minister during the short democratic period in 1958-60), the two districts of Kanchanpur and Kailali formed one constituency in the 1958 election. As a result of the population pressure, Kailali alone was divided into four constituencies in the 1994 election, and the district consists today of 43 different village development committees. The total population of Kailali is about 410,000 (Harka Gurung 1996).

Many Pahaaris from the surrounding districts of Dhoti, Dadeldhura, Baitadi and Achaam have settled permanently in Kailali and Kanchanpur. Some came on their own initiative, but many were resettled by the government. One of the main objectives of the national resettlement department was to resettle landslide victims and others who had been affected by disasters (Gurung 1997).

Indians who originally were invited by the Nepalese government to settle and develop this "remote" area, still form a major part of those who live in the bazaars. From being the only businessmen in the Tarai, they have now been followed by many Pahaaris. Until a few years back, Hindi was the vernacular taught at school, and even today Nepal TV is not available west of Butwal. Those who have TV, watch the Indian national channel Doodarshan or the Singapore-based ZTV.

Dhangadhi is the administrative centre of the Karnali zone and the main bazaar in Kailali. Dhangadhi has grown rapidly during the last two decades. 25 years ago, there was no bazaar in Dhangadhi, only a couple of shops. Now, the Main Bazaar, a road of approximately five-six kilometers, is full of shops. There is a great variety of luxuries available, for instance Kenwood mixers, Japanese cameras, coffee sets as well as foreign chocolate. These "luxuries" are not for the local peasants who produce at mere subsistence level, but Dhangadhi has become an important border town for Indians who want to purchase foreign luxuries at lower price.

The Main Bazaar of Dhangadhi consists of one long road which ends at the crossing Chauraha. At Chauraha, the road either goes south to the Mohana border and further into India, or it continues northwards to Attariya, where it joins the East-West Highway.

The Tharus of the Far West

The Dangoras and Ranas form the two main Tharu groups in the far west. There is also a third group known as Kunna. Neither the Dangoras nor the Ranas commonly refer to each other as Tharu. The Dangoras, like many of the Tharu groups in the eastern and central part, call themselves Chaudhary, and they are also commonly referred to as Chaudhary. The Ranas, on the other hand, call themselves Rana, whereas the Kunnas also commonly use the name Chaudhary. The Kunnas are often considered as a Dangora subgroup.⁽⁴²⁾ The Kunnas themselves claim that they are a separate group, and they make a distinction between themselves and the Dangoras. "We Kunnas only marry other Kunnas", they told me, and emphasised that they worshipped different gods than the Dangoras. Usually, Dangoras and Kunnas would refer to each others as Chaudhary. In some contexts, however, they would make a distinction between them.

The differences between the Tharu communities are significant, both when it comes to their languages, religious and cultural practices, as well as to phenotypical signs, such as dressing and house styles.

The Ranas usually live in small houses in villages which are more dispersed than the Dangoras. The Rana house, which is a one-storeyed building, faces east and has a "mezzanine", where containers of grains as well as other goods are stored. The framework is made of wooden pillars, and walls are made of a particular type of grass, khariya, tied to the pillars and later plastered with mud usually made of clay, cow-dung, water and chaff of wheat.

The Rana women frequently decorate the walls with bas-reliefs painted in colours and adorned with pieces of mirror. Large earthen granaries, known as kuthias, serve as inside walls and divide a house into rooms. The kuthias are approximately the height of a human. The inner part of the house is restricted to members of the family, and normally no stranger is allowed to enter. The kitchen rosaia is placed in this inner part of the house, and the worship of the deities is taken care of in the kola, a separate room next to the kitchen. The outer parts are for sleeping purposes. Every house has a verandah which opens out into a courtyard where much of the activity takes place. On the doorstep of a house, there are some raised platforms of mud for the household deities.

The Rana female dress is very bright and colourful, and consists of a skirt (gangriya) and a bra-like blouse (angiya). The skirt has a beautifully embroidered back part (got). On the head, the women wear a black shawl (urnia), and around each of their ankles, they wear two big anklets (paela). Old women will usually tie their hair very tightly on top of their head, whereas the young women wear their hair in two braides on each side of the head. The rest is tied in a top at the back of their head.

The Rana village headman (balemansa) was originally the link with the government. He was appointed annually during a village meeting (kacheri) by all the heads of the households. The village headman was usually appointed from the better-off section in the village, and it was a prestigious position. When guests came to the village, they would normally eat and stay with the headman, who was a person who "had enough to eat". He was also entitled to have a few days of free labour from all the houses every year. If the villagers were pleased with their headman, he could retain this position for many years. The headman was helped by a village watchman (chaukidaar) who was also appointed annually during a meeting. This was often a poor person, "one who did not have enough to eat", and he was paid annually by all the households in the village. The watchman was sent around to tell the villagers about unpaid community work (begaari), meetings, or to collect money for the collective village rituals. Guests could also sometimes stay with the watchman. While the headman was a "big person" (thulo manchhe), the watchman was a "small person" (sano manchhe).

The headman was not a religious leader, because rituals were taken care of by a specialist (bharra). Hasan (1992), who uses the term gunthera, describes this specialist as "a multipurpose spiritual caretaker". There may be many bharras in each village, and they function both as healers and shamans. Each village often has a main bharra, who is responsible for the main offering to the village deities.⁽⁴³⁾ The villagers in Geti would refer to this bharra as the thulo bharra, "big bharra". This big bharra often has several villages under his responsibility, and is therefore not a resident of the village. I will hereafter use the term big bharra to distinguish him from the ordinary village bharra whom I hereafter will refer to as healer/shaman.⁽⁴⁴⁾

The big bharra was also appointed on a yearly basis. In the village where I did my community study, for instance, a big bharra was not re-elected because the villagers were not satisfied with his work. Neither the headman nor the healer were inherited positions, and they were divided between several persons. The headman would decide when the various festivals, such as Holi and Diwali, should be celebrated, as well as the time for the collective village ritual. After the headman had decided the time, the watchman would tell the villagers, and if necessary collect the items or money needed. The big bharra would finally be called for to perform the ritual.

The Dangoras traditionally lived in beautiful longhouses in rather compact villages (cf. McDonaugh 1984b; Krauskopff 1989a). One longhouse was normally inhabited by senior parents, their unmarried children, their married sons and the sons' wives and offspring. From Dang-Deukheri, often whole villages migrated westwards, and these were re-established in Kailali and Kanchanpur. Today, this migration pattern is still visible, my Dangora informants told me. And Dangora villages in the far west will often differ, depending on whether they originally came from Dang or Deukheri. The languages and traditions of two neighbouring Dangora villages may be slightly different, because one of the villages originally came from Dang, while the other came from Deukheri.

The houses are like the Rana house, one-storeyed and made of wood plastered with white-painted mud. The walls are often decorated with bas-reliefs made by the women, with designs of various animals like elephants and peacocks. The roof is wide and thatched with grass, and the rows of containers, called deeris, divide a house into separate rooms. These deeris are placed on both sides of a corridor and they divide the house into individual sleeping quarters for the various nuclear families. (See McDonaugh 1984b; Krauskopff 1987a.) More than 50 persons could live in one and the same house, and eat from the same kitchen. A Dangora village is usually placed in a north-south direction, and the houses are also organised according to a north-south direction. The kitchen, as well as the deity room, in which the tutelary and ancestral gods of the patriline reside, are normally placed in the north-eastern part of the house.⁽⁴⁵⁾

The appointed Dangora village headman is called mahaton, and he would normally be the head of the village assembly (khel). The assembly is composed of all household heads in the village. The headman was originally the link between the village and the government (sarkaar). Similar to the Ranas, the traditional position of the mahaton has now been replaced by the formally elected ward chairman (adhyaksa) (McDonaugh 1997). Unlike the Ranas, however, the Dangoras have hereditary priests. The tradition among Dangoras in Dang was a system in which political and religious powers were centralised in one and the same person, the desbandhya-gurwa. He was at the same time priest, magician and chief, and Krauskopff calls him "perfect magician". The political power was not distinguished from magical power, and for the Tharus, Krauskopff argues, the strong human being was the one who could pacify the malevolent forces and guarantee the wellbeing of the country (desa) (Krauskopff 1989b).

According to Krauskopff, the various population groups of a particular desa were united under a tutelary deity.⁽⁴⁶⁾ It was therefore in many ways the god that governed, and "ethnic substance" was of minor importance in what is now conceptualised as the area of the Dangora Tharu. This is also expressed in the most well-known myth of origin in Dang which tells how the valley of Dang was divided between four gods. Each god gained control over a particular mythic territory. Some of the patrilines started to worship the god as an ancestor and claimed right to hereditary priesthood in the area controlled by their particular ancestor. This resulted in what later became the distinction between priestly patrilines (desbandya-gurwa ) and their clients (barin). This structure has now changed a lot both in Dang and in Kailali, where the former so powerful desbandhya-gurwa has been reduced to a healer (cf. Krauskopff 1989b). Even though it is difficult to separate the political from the religious on the Indian subcontinent (cf. Neale 1969; Price 1991), none of the positions among the Ranas (bharra, chaukidaar, balemansa) seems to be of the same religious-political character as the desbandhya-gurwa among the Dangoras. The Dangora social structure is also considered to be special in a Nepalese context (cf. MacDonald quoted in Krauskopff 1989a).

The different dressing style between Tharus in Dang and Deukheri is seen mostly on the female part of the population. The women of Dang usually wear saree blouses over a white cloth, lungi, which is worn like a skirt down below the knees. The Dangora women also used to tattoe their legs from the knees down to the heels, as well as their arms from the elbows down to the wrist. In Deukheri, the women often wear colourful skirts known as lahenga and a sleeveless blouse, kurti. The women - both in Dang and Deukheri - tie their hair in a top at the back of the head. Today, the saree, the most commonly worn dress among Nepalese women, is worn by some of the married Tharu women.⁽⁴⁷⁾ The girls who go to school will wear skirts and blouses, as well as the popular kurta surwal, which is a dress worn over baggy trousers.

The Pahaaris are considered to be much more hierarchically oriented than the Tharus, who, as I pointed out in the introduction, have commonly been described as a "tribal" people. One of the criteria given "tribes" in order to distinguish them from castes, was the egalitarian nature of "tribal" society (cf. Introduction). There are, however, both "caste"-like and "tribe"-like aspects of Tharu society, and like most of the societies in South Asia, Tharus are organised according to hierarchical principles. As I will discuss in Chapter 3, the Dangoras in Dang, for instance, differed between priestly and non-priestly patrilines (gotyar) (cf. Krauskopff 1989a). The Ranas, on the other hand, were subdivided into hierarchically ranked subgroups known as kuri (cf. Srivastava 1958).

The subjective identification as Tharu varies quite a lot among the Tharus in Kailali. The way the Tharus refer to themselves and other Tharu groups was not at all congruent. Both Ranas and Dangoras would use the term Tharu/Tharuni to denote human beings, as well as husband and wife. Sometimes the term Tharu was used as a collective denominator, but this was always vis-à-vis the Pahaaris. I will come back to this in more detail in Chapter 4.

Some Ranas claimed that they are Rana Thakur, and not Rana Tharu, and that the term Tharu is a misnomer of Thakur. "We Ranas are not Tharu, but Rana Thakur", many Ranas said, and one of my Rana informants told me that the Ranas had been kings (raajas). Originally they came from Rajasthan in India and were of Rajput descent. Ranas therefore, he explained, had originally not consumed meat or alcohol because of their high Rajput status.⁽⁴⁸⁾ Srivastava (1958) also mentioned that some of the higher Rana subgroups claimed a Rajput (Thakur) origin, and that they "lost caste by using intoxicating liquor and raising fowls" (ibid.:14). Srivastava called this claim for Rajput/Thakur status a process of "Thakurisation". Some Dangora Tharus also claim Thakur background. One of my Dangora informants told me a story similar in nature about the Dangoras' original Thakur status. This person, who is well-educated and has studied pharmacology in Kathmandu, is now running a medical store in a village in Kailali. According to him, Dangoras were of Thakur status and descendants of King Ram Chandra of Ayodhya. As Thakurs, they had worn the holy cord (janai) which is worn by the twice-born castes.⁽⁴⁹⁾ Because of the Muslim conquest and their onslaught on the Hindus, the latter had stopped to wear the holy cord and later escaped through the jungle to Dang. They escaped on horseback, and that is why the Dangoras worship the horse as a deity today. In order to stop the Muslims, they placed a piglet in front of their house, and the Muslims, who consider pigs impure, stopped persecuting them. Since that time, he pointed out, Dangoras started to eat pork, something they had not done in Ayodhya (eating pork is onsidered a low-caste practice). Later, the name Thakur changed and became Tharu (Thakur bata, Tharu bayo). The distinction made between kingly/noble lineages and commoners, is one of the criteria common in a caste system (cf. Introduction).

What these two narratives show, is a tendency which has been rather common among people on the subcontinent (see Srivastava 1958; Sinha 1962; Srinivas 1996). Historically, the Kshatriya was an open and accomodating categoory that recruited from a wide variety of castes (cf. Srinivas 1996:75). A necessary concomitant, if not a precondition for succeeding in a claim to Thakur status and "kshatriyahood", is, according to Srinivas, "Sanskritisation" which he explains as "the acceptance of the rites, beliefs, ideas, and values of the great tradition of Hinduism as embodied in the sacred books" (ibid.:75). The Great Hindu epics have played an important role in the spread of a Sanskritic Hinduism, and what Srinivas considers a common culture throughout India.

Although Tharus to a great extent have been and still are illiterate, and thereby have not been able to read these "sacred books", the process of "Sanskritisation", as Srinivas also stresses, is a many-sided cultural process. Among Tharus in the far west, public readings of the Vedic literature and the Great Hindu epics (Mahabharata and Ramayana)⁽⁵⁰⁾ by pandits, or what Srinivas calls "trained masters of the art", was a very popular pastime. These public readings, known as kathaa or harikathaa, have made it possible for Sanskritic Hinduism to reach even the illiterate masses (cf. Srinivas 1996). In my village of residence, for instance, such kathaas were commonly arranged by independent households. To arrange a kathaa is considered prestigious, and since it also is a great expence because of items needed (fruit, sweets, incense, tread etc.), it was mostly the better-off villagers who had the means to do so.

Pahaari is a term commonly used about people from the hills. The Pahaaris, however, consist of people from different districts and of different jaats, who do not necessarily have much in common. It is particularly the twice-born castes Brahmin, Chhetri and Thakuri the Tharus think about when they refer to anything connected with the term Pahaari. Tharus will, for instance, refer to their customs and their language as Pahaari calan (Pahaari customs) and Pahaari bolii (Pahaari language).

I will briefly describe some of the main differences between Pahaaris and Tharus. The houses differ in both colour and building style. The Pahaari houses are often two-storeyed buildings and brown- or sometimes red-painted. The dressing style also differs, although this is not so easily seen from male dressing, since all males now tend to dress in "western clothes", trousers and shirts. The typical Pahaari male dress, however, consists of a kurta pajamas, often white in colour, over which a black vest is worn. On the head a cap (topi) is usually worn, similar to the type Nehru always wore. Pahaari women wear mostly saree, and when a woman is married, she will put red powder (sindur) into the parting of her hair. The twice-born Pahaaris from the far western hills are considered more orthodox than their counterparts in the Kathmandu valley and the central part of the country. This orthodoxy is related to Hinduism and its rules about ritual purity and impurity.

The high-caste women live a very secluded life and stay mostly at home. They are not allowed to walk around freely, and one hardly ever meets them alone in the village.⁽⁵¹⁾

They often pointed out to me that Tharu women had much more freedom than themselves. "If we walk around alone, people will gossip immediately. We have to stay inside", they pointed out. During their menstruating periods, they are considered impure and are not allowed to cook or enter the kitchen. There is often a separate house where they will sleep during their periods. Many women take off their blouse and make the food only wrapped in the saree cloth (doti) in order to be as ritually pure as possible.

In front of a Pahaari house, there is usually the sacred tulsi plant, which is worshipped regularly by the women in the household. Pahaari men of twice-born status often wear the thread known as janai which indicates their status as twice-born and even during the cold winter months, the twice-born men will take a bath early in the morning.

Many Pahaaris, although they are born in the Tarai or have lived most of their life in the Tarai, still identify strongly with their village in Pahaar, which they will refer to as "our village" (hamro gaon) and our house (hamro ghar). Their house in Kailali is also talked about as home, but the identification with the village in Pahaar is, as far as I could understand, very strong.

The Tarai has a past characterised by migration and flux (cf. Chapter 1). Despite this instability and flux, villages tended to be mono-jaat villages, in the sense of ethnic group. Today, however, Ranas, Dangoras and Pahaaris often live in the same villages. This is also the case with Geti, the village I moved into.

Before I go on to describe this village in more detail, I will just give an example of the fluctuation and instability of some of Geti's neighbouring villages, as it was told by one of the villagers.

30 years ago, there were only Dangoras settled in Junapur. Maybe one Pahaari, but not more. Today, on the other hand, there are many Pahaaris in Junapur, whereas in Pakalpur, the village next to Junapur, there are no Pahaaris. Since a Pahaari once died in Pakalpur, Pahaaris believe that the place is haunted (bhuut lagcha) and therefore they do not want to settle there. Pakalpur, however, was a village earlier inhabited by Kunna Tharus. The Kunnas sold all the land to Dangoras and moved to Kanchanpur. In Shreepur south of Geti and next to Pakalpur, there have always been Pahaaris who were big landowners (jamindaars). In Baskera village north of Shreepur, there were mainly Ranas, but nowadays, there are Pahaaris and Dangoras settled in Baskera. The Ranas left for Kanchanpur or India.

It is now about time to move into Geti, the village which became my home for the months I spent in this "remote" corner of Nepal. Geti is situated about 12 kilometers northwest of Dhangadhi and is a rather large village. The village is not far from the main road between the Indian border and the East-West Highway. Attariya north of Geti is the crossing where the Highway continues westwards to Mahakali river or towards Karnali river to the east.

Adjacent to Geti is the Geta eye hospital, a Norwegian Church Aid-funded hospital, which to a great extent has recruited its staff among the locals. Next to the local patients, people from the neighbouring hills as well as Indians come to the hospital. Transport to and from Geti village is good: buses and trucks, as well as tongas (two-horse-driven carts) and tempos (small diesel-driven vehicles) run frequently. So do cycle rikshaws and dunlops (buffalo carts).

The next chapter is my attempt to give an ethnography of Geti, where I will describe the people as well as the economic resources of this particular village.

Chapter 3 Geti Village

Throughout my thesis, I will explore the various forms of Tharu social identification, and the contexts in which these are expressed. In order to do so, it is necessary to relate them to the local social structure.

In his thesis, based on a study of a village in Dang, McDonaugh (1984a) mentioned three important pillars in Dangora Tharu society. These were (i) the village (gaon), (ii) the house (ghar) and (iii) the patriline (gotyar). These pillars, or units of identification, were so closely related to each other that McDonaugh found it difficult to dicuss them separately. Among the two main Tharu groups in Kailali, Rana and Dangora, these units were also important as identity markers. A village is made up of people who belong to houses (ghar) and patrilines (kurma/gotyar), and I will discuss them all in detail.

Before doing so, however, I will give a description of the village, focusing mainly on three issues: (i) the economic resources of the village, (ii) its various jaats and (iii) the village administration.

A "Bird's-eye View" of Geti

Most of the villagers in Geti are peasants (kisaan) and practice a subsistence farming, if possible with some of the surplus for sale. Some of the twice-born males are employed in service industries, usually office work of some kind, either in Attariya or Dhangadhi. Some of the villagers in Geti (Dangoras and Kunnas, no Ranas) work at the hospital.

In order to be counted as a kisaan, one has to own land. Only the kisaans are represented in village meetings. Not everybody owns or has enough land, and they may therefore often "sign" tenancy contracts of various forms. As long as a person owns some land, however, he is counted as a kisaan. Tenants are commonly referred to as kamaiya. A kamaiya is usually considered part of his kisaan's wealth (cf. Rankin 1997), and therefore they are not represented in village meetings. Kamaiyas may either live in separate houses (bukhraha) on their kisaan's land, where they cook their own food, or the kamaiyas live and eat in the kisaan's house. The difference between kisaan and kamaiya, as practiced among most of the Tharus, is not strongly marked by class distinctions (Rankin 1997). A kisaan can soon end up as kamaiya for some time, whereas a kamaiya can get hold of land and become a kisaan.⁽⁵²⁾ Among those engaged in various kamaiya contracts in Geti, the Dangoras form a majority. At the time of my stay, no Ranas in Geti worked as kamaiyas. ⁽⁵³⁾

Illegal settlements of landless people is a problem in Nepal, and squatters are commonly known as sukumbaasi.⁽⁵⁴⁾ Outside Geti, on land owned by the eye hospital, there were 11 illegally settled sukumbaasis. The majority of them were Kunna Tharus. None of the villagers in Geti were squatters, however. People were reluctant to give information about the exactly amount of land owned, and I do not have a record over landholdership in Geti. The way people talked about each other gave me nevertheless some ideas. Land is usually owned jointly by the male members of a house. The two brothers in the Rana house I lived in, owned jointly seven bigha. They were by others in the village considered as better-off people. They had a surplus for sale and could wait until market prices were high before selling it. When a house splits, which is a common phenomenon, land is also divided. As a result, there are today many kisaans with only one or two bigha each.

There are only three big landowners (jamindaars) in Geti, and they each own 25-40 bigha of land.⁽⁵⁵⁾ Their big Indian-styled bungalows are easily distinguished from the rest of the houses in the village, which are mainly made of mud. The landowners in Geti are either Brahmin or Chhetri. The relation between kamaiyas and jamindaars is often quite different from the kisaan-kamaiya relationship described above. The system of debt bondage occurs more frequently, and the kamaiya may owe the landlord several thousands rupees. In Geti, for instance, a Dangora family owed their landlord 40,000 rupees. During my stay in the village, the kamaiya family ran away from their jamindaar and although he contacted the police, they were not found. For a highly indebted kamaiya to change jamindaar, he has to find a new one who is interested in his labour and therefore willing to pay his debts to the former landlord. By paying off a kamaiya's loan, a jamindaar "buys" cheap labour which is used in his market production. The debt is inherited and passes on from father to son. This is the kind of bonded labour which BASE is fighting against (cf. chapters 1 and 5).

Share-cropping is the most common form of tenancy, and in Geti the practice known as adhiya adhiya (lit. half-half) is the most common, dividing the harvest equally between the tenant and the landlord. A leasing system, where tenants sublet and cultivate a fixed plot of land for a certain period, is also practised. Babu Ram Chaudhary,⁽⁵⁶⁾ for instance, practices adhiya adhiya with one landowner and leases land from another.

According to some landless Tharu informants, parts of the land in Geti are owned by non-residents (bhairako manchhe). I do not know whether this is true or only reflects frustration expressed by those who do not own land. The system of absentee landowners has been a problem in many parts of the Tarai (cf. Chapter 1).

The main crops grown are rice, wheat, mustard seeds and various kinds of lentils. It is common to distinguish between dry-field cultivation (kharif) and wet-field cultivation (rabi). The main dry-field crops are wheat, maize, mustard, pulses, whereas rice is the main wet-field crop. Some of the jamindaars practice cash-crop cultivation. Sugar cane is a highly income-generating source, and demands a well-functioning irrigation system as well as fertilizers (Cederroth 1995). Although Geti has an irrigation system which functions well,⁽⁵⁷⁾ sugar cane is not commonly grown. The villagers told me that the refining factories were too far away, and that this made it less attractive to cultivate sugar cane. Wild fruits and honey are also gathered, mostly for own use but sometimes also for sale. Fishing is a popular activity and fish an important resource. Fish (maachaa) plays a central role in some of the Tharus' ritual activities (cf. Krauskopff 1987b), and some people will also sell their catch. The vegetables grown are mainly for own consumption, although a few people in Geti sell their vegetables to the small shop at the hospital. Those with milk-giving buffaloes also regularly sell milk to the teashop at the hospital and thereby earn a few rupees.

Tharus are known to be "big rice eaters", and rice is the most important crop. Since we do not grow rice in Norway, they all wondered what on earth we would eat instead of rice. Hearing that we eat a lot of bread (roti), they were not very tempted to visit Norway. Although Tharu women make the most delicious roti, of corn, riceflour and wheat, roti is considered a poor man's food and most Tharus will eat it only in addition to their rice. Rice production is both labour- and water-consuming. Rice is usually sown in late April, and weeding is done continuously until the monsoon starts - usually in June. During the monsoon, the seedlings are transplanted, and the rice harvest takes place in September-October.

There is no forest covered area where fuel and fodder can be collected within the boundaries of Geti village. The villagers therefore usually go to a forest-clearing nearby the hospital and, as far as I was told, this was part of the hospial's property. The collection of fuel for daily use is a time-consuming job, and it is usually a task for girls. As already mentioned, cutting trees has been forbidden the last 15 years. Illegal cutting, however, occurs quite frequently during night-time.

In order to avoid trouble from the armed forest guards (bal bhutti), bribing is common. Each household head in Geti payed 60 to 70 rupees to the village headman who then would pay off the forest guard. During my stay, one of the villagers in Geti was caught by the forest guards and jailed for illegal cutting. His family went around to borrow enough money to have him released, and this cost them 6,400 rupees.

The building of a traditional Tharu mud house demands a lot of wood. Lack of wood has therefore made brick houses (pakki ghar) more common and popular. The traditional mud house was usually built with help from neighbours and relatives, who would be given meals and liquor (raksi) in return. To build a brick house, on the other hand, requires skilled labour and is therefore more expensive. To finance a brick house, many take loans (rirn) from the Indian moneylender (mahajaan). Although his interest (35 per cent) is much higher than the interest on a bank loan, people preferred the moneylender to whom they may pay in kind.⁽⁵⁸⁾

Next to agriculture, animal husbandry is a vital source for the Tharus (see Krauskopff 1986). Poultry raising is a female activity and used for both own consumption as well as for sale. The money earned on the sale of eggs and chicken, is usually kept by the women themselves. Goats, pigs and sheep are also frequently raised, and the meat is eaten on special occasions, for instance during the main festivals dashain and diwali (see below). The Ranas keep their animals in separate sheds (saar) in the courtyard, whereas the Dangoras keep their animals in the south-western part of the longhouse. Chicken, however, are usually kept inside the house both by Ranas and Dangoras.

Cattle is of vital importance both in the agricultural activity - as a means of transport - as well as for the production of dairy products, such as milk, curd (dahii) and clarified butter (ghee). A pair of bullocks are needed for plowing the fields, treshing the crops, as well as for pulling the buffalo carts (dunlops). The water buffaloes (bhaisii) may also give milk, although a number of cows are usually kept for milk-giving purposes.

Village life is influenced by the agricultural cycle, and this reflects the religious and ritual celebrations as well. The Tharu annual cycle starts on the first day of the lunar month magh (usually mid-January),⁽⁵⁹⁾ with the celebration of Tharu New Year (magh sankranti).⁽⁶⁰⁾ During this "New Year" many people, both Tharus and Pahaaris, take a ritual bath at the confluence of two rivers. Magh is also the time when new contracts between kamaiyas and kisaans are settled. Except for the mustard harvest, magh is not among the peak agricultural periods, and the time-consuming Rana weddings are celebrated then. The appointment of a headman during a village meeting also takes place during this month.

In phagun (February-March), the harvesting of pulses starts, and the Rana girls are getting their skirts (gangriya) and blouses (angiya) ready for the Holi celebration. The traditional Rana celebration of Holi starts on the full-moon day of this second month and goes on for a whole month. The village girls who are married but childless, go home to their natal home to "play Holi" (holi khelne). Marriages among Dangoras are often celebrated in phagun.

The harvesting of pulses continues in chait (March-April), and during this month, the young Ranas finish the Holi celebration by "sending Holi away" (holi partaune). The Rana and Dangora women celebrate a festival known as carai, and chait is also a month Ranas use for their betrothal negotiations (magni).⁽⁶¹⁾ These negotiations continue in baisakh (April-May), when the Nepalese New Year (naya barsa) is celebrated all over the country. A second carai is celebrated by the Tharu women in this month as well, and for this second carai, the married Rana girls return to their husbands' houses.

Before the monsoon starts, usually during the month of jeth (May-June), the sowing of corn takes place.This is the beginning of the busy agricultural season which reaches its peak with the transplanting of rice. In aasaar (June-July), the big bharra usually takes care of the annual ritual (aasaari pooja ) in the village shrine. After the monsoon, in the month of saaun (July-August), the time is ideal for fishing, and people catch fish in the fields, in ponds and in rivers. A festival, tiij, is also celebrated by the women. The corn harvest usually starts in bhadaun (August-September), when the Dangoras celebrate Krishna's birthday, called astimki.

The main festival in Nepal, dashain,⁽⁶²⁾ which is equal in importance to our Christmas, takes place during the bright fortnight of the month aashvin (September-October). Dashain is celebrated for several days by Dangoras and Pahaaris, whereas the Ranas do not celebrate Dashain. In kartik (October-November), another main festival, known as diwali, is celebrated.⁽⁶³⁾ The rice harvest goes on, and vegetables, such as garlic, tomatoes, potatoes and cauliflower are sown, as well as wheat and pulses. The sowing of mustard takes place in mangsir (November-December), and in the rather quiet month push (December-January), the Rana wedding season starts.

Geti was originally a Rana village, but because of the migration which has taken place during the last years, it is today a village inhabited by many different jaats, both Tharus and Pahaaris, including a few Kunnas and the various occupational castes. The occupational castes were present in Tharu villages before the main Pahaari migration started. As pointed out in the Introduction, they belong to the Pahaari category, although the label Pahaari is mostly associated with the twice-born castes; Brahmins, Chhetris and Thakuris. (See also Chapter 1.) The occupational castes do not usually own land, but make a living out of their manual work, as blacksmiths (lohaar), tailors (damai), menders (misteri) and goldsmiths (sunaar).

The settlement pattern in Geti is to a great extent jaat -based. To stay near one's own jaat (afno jaat) has been, and still is, preferred among the villagers. Geti is therefore a village of many different clusters or neighborhoods (tol), which do not necessarily have much to do with each other. Geti is divided into two administrative units known as wards, and each ward has an elected chairman called adhyaksa.⁽⁶⁴⁾ There were 118 independent houses (ghar) in the village. Out of this number, the Dangoras formed a majority of 50, whereas the Ranas numbered 29 and the Pahaaris 37.⁽⁶⁵⁾ There were only two Kunna houses in the village.

The Rana neighbourhood is located in the southern part of the village. The Rana houses, described in Chapter 2, face east and are grouped in compounds of patrilines, kurma. South of the Rana neighbourhood is the village shrine bhuinyar, where the village deities are enshrined.⁽⁶⁶⁾ The bhuinyar is a Tharu institution, and the Pahaaris in Geti do not participate in activities taking place there.

Dangora houses are found all over the village, except in the middle of the Rana neighbourhood. Many Dangoras are settled in the north-eastern part of the village, and the Dangora houses are also commonly clustered around the patriline (gotyar). Except for one house, all the Dangora houses in Geti are placed in a north-south direction, as is common practice among Dangoras. The one exception in Geti, placed in the opposite direction, belongs to a patriline known as ultawa (lit. opposite). The two Kunna houses in Geti are inhabited by two brothers, and they are placed next to each other in the northern part. A few Rana houses are placed separately in the north-western part of the village. The reason why they are separate (allak allak) is that other Ranas who lived in this part of the village before, had sold their land and moved.

The occupational jaats also stay near to each other. The house of the blacksmith, for instance, is next to the house of the tailor, who again is the neighbour of the village mender. Although they bought land where it was available when they first moved in, the twice-born castes have grouped around their own caste as well. Near to the Rana house which became my home, there were two Brahmin houses, whereto both of the families had migrated from the same place in Dadeldhura district. The cluster settlement among the Pahaaris is very clear in a part of the village newly settled by migrants from Achaam district which has virtually become an Achaam neighbourhood. Since this part of the village is a bit away from the village core, many Tharus did not know much about these newcomers.

The linguistic situation in the village is rather complex. The three main languages spoken - Rana, Dangora and Pahaari/Nepali - are, at least with some goodwill, mutually intelligible. The Rana language is similar to the Hindi dialect Awadhi (Bista 1967), whereas the Dangora language is similar to Bhojpuri (Krauskopff 1989a). The Pahaaris would mostly speak Nepali, but they had also their local hill dialects, which I had much greater problems in understanding than for instance the Rana language.

Nepali is the lingua franca in today's Tarai (cf. Population Monograph 1995). The younger Tharu males in Geti would therefore in general know Nepali. They had often spent a few years at school where Nepali is the vernacular taught and spoken. Since the men move more around than the women, they are in contact with Nepali-speaking people, which makes their skill in Nepali good - even for those who have never been to school. Less than twenty years ago, informants told me, Hindi had been the vernacular taught at school, and those who had been to school at that time, had learnt to read and write Hindi. In most interaction between Ranas and other jaats in Geti, Rana would be the language spoken. Both Dangoras and the Pahaaris who had lived in the village for a long time, would turn to Rana in such encounters. This made Rana into the lingua franca of the village.

The focus of my thesis is on the Tharus, and I therefore spent more time among them than among the Pahaaris. Approximately 70 per cent of the inhabitants in Geti are Tharus. The village is therefore still much in their hands, which is apparent in many aspects of village life. The Pahaaris in Geti do not participate in the collective ritual activity in the village shrine. This exclusion is not only related to ritual matters, but also to more political ones. As pointed out in Chapter 2, on the Indian subcontinent it is difficult to distinguish strongly between political and religious matters.

Today, the traditional Tharu organisational forms (cf. Chapter 2) have to a certain degree been replaced by the formal Nepalese system. However, in Geti, like many other villages in the far west, the traditional system of an appointed headman coexists with the formally elected leaders. The importance of the traditional headman and his functions may vary from village to village. In Geti, the formerly elected ward representatives (adhyaksa) have to a great extent taken over the role of the headman. The duties which the headman traditionally had taken care of, were now in their hands. The headman nowadays had less power, and therefore this position had become less prestigious.

According to Basu Dev Chaudhary and Birbal Rana, the two ward chairmen, the headman position had become a very difficult one. "It is a lot of work, but no respect from the villagers", they both emphasised. When the headman tells people to join for community work (begaari), people do not come. Many are also reluctant to pay the yearly amount of rice, known as tihai to the headman. From what they said, it seemed as though the former so prestigious headman position had become more like the duties of the watchman described in Chapter 2. One of the watchman's duties was to be the village messenger. In Geti, the positions of headman (balemansa) and watchman (chaukidaar) are now held by one and the same person.

Each of the two wards in Geti has a committee of four or five members (sadasya). The head of the committee is the ward chairman (adhyaksa), who represents the ward in the village development committee (VDC). The VDC is the lowest administrative level in Nepal and was introduced after the partyless panchayat system ended in 1990 (cf. Chapter 1). Each VDC is made up of nine wards and headed by a chairman also called adhyaksa, like the head of the wards. Geti, together with six of its neighbouring villages, form the Geta village development committee (Geta Gaon Samitti). The VDC adhyaksa is elected from the Congress Party, and he is a Brahmin and schoolteacher from Geta village. Both the ward chairmen as well as the members of the two ward committees in Geti are Tharus. The two ward chairmen in Geti are elected from the Congress Party.⁽⁶⁷⁾

The Significance of the Village

In most of the literature about the Tharus, the village unity has been emphasised. Now that villages consist of several different jaats, which to a great extent stick to themselves, the concept of village (comm)unity becomes more problematic. This brings me to a brief outline of the central role village studies long played among scholars who were working on the subcontinent.

The village community was considered a political unity and often depicted as self-sufficient "little republics" and "almost independent of any foreign relations" (Metcalfe 1830, quoted in Dumont 1970). They were conceptualised and idealised as isolated, never-changing entities (cf. Munro 1812).⁽⁶⁸⁾ This view has been challenged, and Dumont and Pocock (1957) argued that the village itself had no sociological relevance. Villages were, according to them, better understood in terms of the caste system and in terms of the relation between dominant and dependent castes. The Dumontian view of the hierarcical caste system as the key concept for understanding India, has in turn been thoroughly criticised (cf. Raheja 1988; Appadurai 1986; Srinivas 1996).⁽⁶⁹⁾ Bailey (1959) disagreed with Dumont and Pocock and argued that "sociological reality is not proved or disproved by the presence or absence of 'solidarity', nor of feelings of loyality. 'Sociological reality' pertains to social relations" (ibid.:94).

In my thesis, I focus on the contemporary situation of Geti village, and I will also argue that the village has sociological relevance. It is a place the villagers refer to and identify with and in which they engage in social relations. Srinivas (1996) has tried to show that the village is not only an architectural and demographic entity, but also a social entity to which the villagers have some loyalty.

One of the first questions Tharus and most rural (and also urban) people in Nepal will ask each other when they meet for the first time, is which village the other comes from. In inter-village matters, village identity becomes more relevant than identities related to castes, patrilines and houses. It seems to be very much the same with what Srinivas has pointed out about rural Indians:

People in and around Geti are also often talked about in terms of the village where they live, as when my friend Tove, who lived in Manhera village, was commonly referred to as Manhera-wari (the one who lives in Manhera).

An important kinship unit among the Dangora is called gotyar, a word stemming from the Hindi word gotra and commonly translated as patriline or clan.⁽⁷⁰⁾ The Rana term for patriline is kurma and according to Hasan (1992), it is a word derived from Persian kunba, which means clan. The gotyar/kurma is solely an exogamous unit, based on patrilineal membership.⁽⁷¹⁾ The patriline is not a localised group, and the members may be dispersed over many villages and often do not know about each other. The gotyar, McDonaugh (1994) points out, is not a corporate group. It does not hold land or ritual sites in common, nor is it integrated by an overall lineage structure. The main function of the lineage is to define, in negative terms, with whom one can marry.

The core definition of a patriline is the pantheon of lineage gods (kuldevta). Each patriline is identified and distinguished by its deities. If the gods are similar, one belongs to the same patriline/patriclan. The Tharus, thus, do not express kinship in terms of "shared blood or bone" (cf. McDonaugh 1994), something which is common among other groups in the region. "The agnatic kinship of shared gotyar membership is ultimately anchored in and conceived in terms of a common relationship to shared deities" (ibid.:43). Krauskopff (1989a) has called this a "filiation spirituelle". When marriage relations are negotiated, one of the first things to be checked is this pantheon. In addition to this exogamous character, the patriline may also be seen as a ritual unit in which the members worship the same lineage gods. This has mainly meaning on a local level, because there is no collective worship of these gods, but something each house is responsible for individually. Each house may also add its own individual deities to the principal lineage gods (McDonaugh 1994), as the following examples illustrates:

Birbal Rana in Geti told me how his family had added an extra deity. Many years ago, one of his uncles was imprisoned, and in order to have him released, the family worshipped their lineage gods without any result. His father then asked another deity for help, and after this, the uncle was released. The family therefore added this deity to their pantheon of lineage gods. Another example was given by a Dangora man in Geti, who told me that his gotyar was called Magar Chaudhary. This name they had got after a special event many years ago, when a Magar (a jaat residing in the hills of central Nepal) came down from Pahaar to sell his beer. While he stayed in the Dangoras' house, the Magar died. To die far away from one's own people is considered inauspicious. People who die in such a way, it is commonly believed, may appear as a malevolent spirit (bhuut) causing harm and afflictions (see Fuller 1992). Deification is a common form of dealing with possible afflictions caused by such deaths, and the Dangoras started to worship the Magar as a deity and, since that time, they called themselves Magar Chaudhary. This last example led to segmentation in that they formed a new, separate patriline/patriclan.

Every household has its resident deities enshrined somewhere in the house, and also often in various places outside the house. The Dangora call their deity room deurar, a room which is most commonly placed in the north-eastern part of the house. The Dangoras' deity pantheon usually contains a horse and a poney made of clay. The horse symbolises the deified ancestor of the priests' lineage, with its right to hereditary priesthood, whereas the poney symbolises the ancestor of the house. The horse is thus the representation of ancestors who gained status of masters of the soil - with rights and links to a particular territory The poney, on the other hand, represents ordinary ancestors with no such links to territory. This distinction reproduces a fundamental opposition on the divine plan between priests and clients, and this is expressed in the legend about the division of Dang valley into mythic territories (see Chapter 6). The Ranas' deity room, known as kola, is a small room next to the kitchen. Their deities are physically represented as bumps made of mud (maato) placed on a platform both inside the kola and on the doorstep of the house.

The two cases I described above, which illustrated how a house can add its deities individually, showed that the deities added were linked to the particular house and its experiences. This illustrates in general the role deities play for the identity of a house and its function as a ritual unit. As the smallest of the three main units of identification described, the individual house with its various deities has great importance on a local level. The operative unit of the patriline is therefore, as McDonaugh (1994) has pointed out, the household.

Among the Dangoras, the oldest male person in the house is usually the one in charge of the worship of the lineage gods. Women are not allowed to worship them in their husband's house, because they are still considered as partly belonging to their natal patriline. Among the Ranas, however, it is often the oldest woman in the house who will worship the lineage gods. Contrary to the Dangora women, a married Rana becomes part of her husband's kurma immediately after the marriage.⁽⁷²⁾ She is from then on not allowed to enter the deity room in her natal home. The various male members of an extended joint family⁽⁷³⁾ own the agricultural land together. This is also the case with cattle, household utensils, fishing nets etc. The property is under custody of the household head. He is, however, not allowed to do major changes, such as to sell land, without consent of the other male members. In Geti, there was a "horror story" frequently told me about the head of a Rana family, who, according to the villagers, was a very bad person (karaab). Without asking his younger brother, he sold the land, kept the money for himself and went with his wife and children to India. The younger brother's family was left in Geti with virtually nothing.

Houses are also named, and the names are often derived from the village where the household was previously settled. A house in Geti, which is known as Dhangadhia ghar (the house from Dhangadhi), indicates that the people had come to Geti from Dhangadhi. Houses may also be named after the head of the household, such as "my" Gorya ghar, which got its name from the nickname of a former head. (The same practice of naming houses also exists among the Tharus in Dang cf. Rajaure [1993].)

My informants had often explained that when a house splits, the new houses become kurma. It was, however, not easy to distinguish between a kurma and a household. A compound of several houses, which I thought were separate houses belonging to the same lineage (kurma), were nevertheless referred to as one "house" (ghar). There is thus not only the separate buildings which indicate that a building is a ghar. Among the Ranas, a ghar may be several buildings clustered in the same courtyard, and the members will all eat from the same kitchen. These houses are conceptualised as one ghar (household). The term house then, can denote three different things: (i) It can refer to the physical buildings (ii) it can be used to refer to a household, which may consist of people living in several buildings within a compound and who share a kitchen and (iii) it can be used in the meaning of "home". The household is thus the fundamental social, economic and ritual unit of Tharu society. And there is a hierarchical relationship among the various members of the household. This hierarchy is based on age, gender and seniority (see below). The father is usually the head of the household and, after his death, the elder son will normally succeed him. The head of a household is the family representative in the village assembly and the one in charge of the family budget. Among the Dangoras, the head of the household, who is usually the eldest male, also takes care of the worship of the lineage and households' gods.

The extended joint family, here as in most of the northern part of the subcontinent, has been the ideal. Breaking up of joint families into smaller units is, however, quite common, and many villagers told me that they would like to live separately (allak allak) in nuclear families. When houses split and become kurma to each other, this indicates that the members of each house also cook and eat from different kitchens. In this case, land is also divided.

Srivastava (1958) mentioned a tendency for larger households to break up into smaller units, so this is maybe not a new phenomenon. Quarreling was a commonly given explanation for why houses split. The household head is usually the one responsible for distributing goods to the members of a house, and the members again may often complain over his unjust distribution of rights and duties. Quarrels between brothers, sisters-in-law, or between father and sons frequently occur. If a person wishes partition and his individual share of the joint estate, he has a right to do so, and according to Srivastava (1958), this is similar to the Mitakshara law prevalent in Hindu joint families. And if this partition takes place during the father's time, a share is also reserved for the father.

In Gorya ghar, for instance, the position of household head was given to the second eldest son. Jagan, the eldest son who was angry about this decision, wanted to establish his own household. During this fission, cattle, money and other movables as well as immovable property were divided between the sons. Although the property is ideally divided into equal shares, such partitions are often occasions of quarrels and disputes. Jagan's wife told me that Jagan did not get his equal share of the cattle. Out of the six buffaloes owned jointly, Jagan was only given one. This was because of the greed of the other members, Jagan's wife emphasised. This greed was later punished by the omnipresent god (bhagwan), when several of the buffaloes remaining in the original house, died from a disease later the same year. (Because of the climate, diseases and deaths among the cattle occur frequently.)

Family relations among Tharus, as well as relations outside the family, are to a great extent based on age, seniority and gender. This is a common feature on the subcontinent. Seniority is something else than age, but these two criteria will quite often coincide. Although the wife of a younger brother may be older than the wife of an older brother, it is the age of the brothers which decides the relation between these women. A son has to show respect to his father and mother. He must also show respect to his elder brother and the wife of this elder brother. A wife of the youngest brother in a household has to show respect to everybody.

Except for children, proper names are not commonly used among the Tharus. They prefer to use the kinship term, and specify this term with the position a person has in the hierarchy of the family.⁽⁷⁴⁾ The eldest brother will therefore be titulated as "eldest" (jetho). The second brother in age will have the term "second" (mailo) added to the kinship term, and the term "third" (sailo) will be added to the third brother. Some of the kinship terms specify whether a person is younger or older in age, and this is common among both Tharus and Pahaaris. I will here only mention some Rana kinship terms to illustrate this system. The terms for elder brother is, for instance, daada, whereas a younger brother is called bhaia. An elder sister, on the other hand, is called didii, and a younger sister lalo. These terms are used both in relation to classificatory and real kin. (For a more detailed description of the kinship system, see Srivastava [1958] and McDonaugh [1984a].)

There are also rules of mutual avoidance, which exist among Ranas as well as Dangoras, and which are commonly practiced in North Indian Hinduism (cf. Tyler 1977). There should, for instance, not be any direct or personal contact between a man and the wives of his younger brothers. The wife of Shuklal's younger brother would therefore place Shuklal's food plate on the ground instead of giving it directly to him. She would never address him directly, either.

Tharu women, in general, have more freedom than the twice-born women. But there seems to be an interesting difference between Ranas and Dangoras in relation to female status and identity. A Dangora woman needs much longer time to become part of her husband's patriline than a Rana woman.⁽⁷⁵⁾ A Dangora woman is not fully part of her husband's patriline (gotyar) until her eldest son has married (cf. Krauskopff 1989a; McDonaugh 1994). As already mentioned, the relations between a household and its deities are of crucial importance. Among the Dangoras, it is the men and the hereditarily linked household priest who maintain relations with the deities through the performance of household rituals. The women of the household have virtually no role in these rituals (McDonaugh 1994). Dangora women are not allowed to participate in rituals devoted to the village and forest goddesses, and prior to marriage, girls do not participate in ancestral and household rituals. The women, however, play a leading role in birth and marriage rituals. The Rana women are also the ones in charge of the marriage rituals, but they would also frequently be the ones to worship the households' and lineage gods.

The life cycle of a young Tharu woman, both Dangora and Rana, seems to be similar, however. Until she becomes a mother, a Dangora girl is considered a bathinya (girl) and takes part in the dancing together with the rest of the girls in her village. Once she has a child, she is considered a wife and becomes more linked with her husband's family. She then has to stop the dancing activities. Whereas Tharu men have a permanent and unchanging gotyar identity, the married women change from their natal group to that of their husband. Rana women do so immediately and fully upon marriage, whereas Dangora women do so more gradually (cf. McDonaugh 1994; Krauskopff 1989a). Although a married Rana girl belongs to her husband's kurma after marriage, she does not spend much time in her husband's house until she has a child. She visits relatives and spends a lot of time in her natal home, in a period known as ninharo. She will also celebrate several of the festivals in her natal home, and dance during the Holi celebrations in her natal village. This "freedom"comes to an end when she becomes a mother. After they become mothers, the married Tharu women stay mostly at home.

A woman will normally own the jewellery given by her parents at marriage. It is considered her property (sampatti), and since daughters do not inherit their father this could be seen as their share of the father's property, as well as an economic security. The eldest daughter in Gorya ghar keeps her jewellery in her natal home, where aya looks after it. This was done, aya told me, to prevent her husband from getting control over it, and thereby maybe sell it. The husband was considered a bad person (karaab) who drank a lot of liquor and, when drunk, he would beat his wife and children.

One of the central questions I raise in this thesis, is whether the Tharus are constituting themselves as an ethnic group. Castes and ethnic groups, which both recruit members by kinship and marriage, are different modes of social organisation based on different principles for being the same kind of people (cf. Introduction). In my attempt to come to terms with what a transformation from a caste structure to one based on ethnicity implies, it has been necessary to describe the local social structure. In this chapter I have, therefore, given a description of the village and the two most important units of identification among the Tharus, the patriline and the household. The sociological relevance of these units depends on contexts. An interesting point to stress, is the relationship between people and gods. People who worship the same deities are considered to be of "the same kind of people". This criterion for a common identity seems to prevail over other criteria, such as the ones linked to territory or "shared blood". But they organise themselves, to a great extent, according to what I have listed as caste principles, cf. Introduction.

The purpose of the next chapter is to look closer into forms and contexts of social exclusion and inclusion and the ways these are manifest in Geti village - both within and between jaats.

Chapter 4 Contexts for Social Exclusion/Inclusion

"If we ask questions about ethnicity, then we shall no doubt find ethnicity" (Eriksen 1992:203-204).

Processes of inclusion and exclusion are highly dynamic and context-dependent (cf. Moerman 1965; Berreman 1975; Okamura 1981; Eriksen 1992), and I will in this chapter look closer into the various contexts in which such processes occur in Geti. A separate collective "Tharu" identity was not a very salient identifying criterion for the Tharu villagers. The most important units of identification among the Tharu villagers were not ethnic in nature, but rather locally tied to kin and village and gods as well as ritual status. Ranas in and around Geti, and also some of my Dangora informants, would frequently claim that they were of Thakur origin and they did not usually refer to each other as Tharu (cf. Chapter 2). Some Ranas even told me that "if Dangoras say they are Tharus, then we Ranas are not". Here I will further explore the way Ranas and Dangoras exclude/include each other. What kind of principles do they base their social interaction on and what are the attributes they chose in order to communicate their differences? The issue of boundaries and boundary marking is therefore central in this chapter.

Boundary Marking and Hierarchical Ranking

Ethnic boundaries are often expressions of status identities to which claims of one type or another are attached. The jaat ranking as it was laid down in the Muluki Ain (cf. Introduction) gave a hegemonic status to the twice-born jaats. The interrelationship between jaat and rank in Nepal matched the structures of power and authority, and this has influenced all parts of society, both higher and lower strata.

Among Tharus, Pahaari customs have to a certain extent become a model against which they measure their own traditions and values. The claim to certain status identities made by Tharus in Geti was much in accordance with what Srinivas called "Sanskritisation", the process by which a low caste or tribe or other group "takes over the customs, ritual, beliefs, ideology and style of life of a high and, in particular, a 'twice-born' (dwija) caste" (Srinivas 1996:88). It is very often in order to improve its position in the local caste hierarchy that a group choses "Sanskritisation" (ibid.). That the mainstream values of Sanskritic Hinduism have influenced the Tharus' own outlook is not so strange, and it indicates that they think of themselves in terms of caste principles. What is also interesting, is that each of the Tharu groups in Geti gave each other attributes which were the opposite of Pahaari customs, and something I will call "negative Pahaari values".

The Ranas in Geti told me, for instance, that Dangora customs were the opposite of Pahaari customs, whereas the Dangoras would utter similar statements about the Ranas. These statements I understand as an attempt by each group to give the other group a lower status than one's own. This is similar to what Berreman (1975:85) has called "mutually antagonistic symbolic degradation". Implicit in such a mutual status degradation, Berreman stresses, is an affirmation of the status of one's own group, and a common feature when it comes to relations between competing ethnic groups and competing castes (ibid.). Eriksen, who seems to agree with Berreman, writes, for instance, that "the others are held to represent lifestyles and values which are regarded as undesirable" (Eriksen 1992:43).

Examples of such mutual symbolic degradation in Geti were many, and the Tharus' statements were very similar in nature to the Pahaaris' general view of the Tharus, so I will therefore start by briefly describing the Pahaari stereotyped version of a Tharu.

According to Pahaaris, Tharus were "socially backwards".⁽⁷⁶⁾ Some of the criteria for this characteristic were related to illiteracy, lack of hygiene, bad dietary and drinking habits, as well as improper female dress and behaviour.

People in Nepal will commonly talk about each other in relation to the women and their attire and attitudes (cf. Gellner 1989; see also Unnithan 1994). Gender, therefore, seems to be a useful analytical tool when it comes to the way the various jaats in Geti would communicate differences and similarities. The Tharu women's dress (and in particular the Rana dress) was considered highly improper because it showed too much of the body. That Tharu women roam around more freely, both in the village as well as in the bazaars, was also considered improper behaviour.⁽⁷⁷⁾

Many Pahaari women commented upon the Tharu women and what they considered their lack of hygiene, and they also demonstrated to me how the Tharu women went to the toilet by just squatting anywhere. The Pahaaris in Geti always asked me how I managed with my food in the village. When I told them that I ate in Tharu houses, they seemed very surprised. Because of the rules concerning food and contact, the twice-born Pahaaris would, for instance, not eat rice in a Tharu house. Rice is categorised as kaccaa food, imperfect food which can easily be polluted.⁽⁷⁸⁾ The concern with pollution was, however, not the only reason why they asked. They also wanted to know how I managed to eat the very spicy Tharu food. When I said that this was not a problem at all, they hardly believed me. Tharu food was known to be almost inedible because of all the chilli.

Some Pahaari friends were different and did not talk about Tharus in the usual "Pahaari" way. When it comes to food, however, they too expressed the same ideas. Due to the heat in May, I had been sick and was not feeling well when I met my Pahaari friends. They commented upon my weakness and immediately related it to the Tharu food I was eating. When they met Tove, for instance, they told her about "poor Sita" (my Nepalese name), who did not get proper food in the village, and therefore had become so weak (kamjori) and skinny (dublo). According to recent anthropological writings (Gellner 1989:6), the politically and culturally dominant groups tend to view those who are peripheral and subordinate to them as inverting their own "natural", "correct" way of doing things. These social attributes given Tharus were the opposite to how Pahaaris liked to think of themselves. They were, therefore, important symbols of identity and self-esteem for the Pahaaris, and because of their hegemonic status, they had never been opposed or contested. Tharu-Pahaari interaction very often showed the discrepancy in status and power, where the "symbolic power" of the twice-born became very clear.

Wherever I went, Pahaaris would ask me why I had come down to this part of the country. When I explained that I had come to learn about the Tharus in this part, they often started to give me lectures about Tharus and Tharu culture. The fact that I was together with a Tharu informant, did not prevent them from being "experts" on the Tharus. My Tharu informant would usually remain silent, but after we were alone, she complained about all this talking as typical for the Pahaaris (Pahaari baanii jastai).

When Tharus and Pahaaris met, such meetings were dominated by Pahaaris. During a day of village work in Geti, some Tharu women were sweeping the village lanes, when three Pahaari men came back from the teashop. The Pahaaris stopped and made comments upon what they considered a pointless thing to do. The lanes would be dirty again at once, they quite rightly pointed out. The Tharu women remained silent, but I told them jokingly that instead of just watching, they could help us. "We do not shit in the lane", one of them, the biggest landowner said, and he added: "We use our toilets." The three men started to discuss the need to teach Tharus how to use toilets instead of the lanes. Since the Tharu women still remained silent, I argued that people would go to the fields whenever they needed to defecate, and that nobody, except for some children, would do this in the lanes.

In this case, gender also came in as a criterion. Women usually show men respect and a common way of showing respect is by not opposing the speaker. As a rule, women will not speak in front of males whom they consider superior in age/seniority. In male encounters between Tharus and Paaharis, the same respect was shown by a Tharu, although he could be higher ranked in age and seniority. Even when Tharu men interacted with Pahaari women, the Pahaari women could dominate the situation. When one of the neighbouring Brahmin women came to visit me in my Rana home, Shuklal, the head of the house, sat down with us, and we all started to discuss a recently arranged Tharu culture programme. The Brahmin woman, who strongly criticised the whole arrangement, dominated completely the conversation. Although I knew that Shuklal disagreed with her, he remained silent. Another example I will mention here, took place during a meeting with one of the leaders of a Rana Reform Movement. One of his colleagues, a female Brahmin, came into his office and started to talk about how the Ranas ought to reform their community. Even this well-educated Tharu remained silent and let her do the talking.

In face-to-face interaction, Pahaaris would generally dominate and define the situation, whereas Tharus remained silent. Pahaari-Tharu social interaction was thus clearly marked by asymmetrical power relations. This asymmetry is inherent in the social status the twice-born Paaharis have as high castes, as well as in an inferiority complex generally found among the Tharus. This fits in with what Despres (1975a) has argued, that ethnic phenomena is probably best understood in relation to a system of stratification and to general theories of power.

Rana and Dangora Antagonism

People of common status, on the other hand, may interact more freely and intimately (cf. Berreman 1975). I thought that their common status as Tharu would make Ranas and Dangoras interact more frequently and that this interaction would be more intimate. What struck me, instead, was the seeming lack of interaction between Ranas and Dangoras in Geti, especially among the women. Tharu women, I pointed out in Chapter 3, move around more freely than their Pahaari sisters. But after they have children, Tharu women spend much of their time at home. They are, therefore, mostly in contact with members of their patriline and sometimes their neighbours. Due to the settlement pattern in Geti, the neighbours would often be of the same jaat and even the same patriline. Social interaction across jaat boundaries, therefore, occurred more rarely among the female part of the population.

During daytime, the women were usually too busy at home to visit other houses, and at night they were afraid of the dark and did not want to go far away from the house. Men, on the other hand, moved more around in the village on their way to and from the fields as well as on their way to the forest, market etc. They were more in touch with people of all jaats. Village meetings, however, became the best occasions for me to observe such male interaction, and I will come back to such village meetings later in this chapter.

Since social interaction across jaat boundaries occurred more rarely among the women, their verbal statements became an important source of data. These verbal statements were, of course, a result of their impression management. Statements always have to be related to practice, and I had fortunately the opportunity to do so. The content of these verbal statements are nevertheless interesting in themselves, because they show how the Tharus use the same criteria as the Pahaaris in order to denote a lower status to each other. Educational status, dietary habits as well as the role of the women were criteria given for what Tharus deemed civilised and uncivilised. A well-educated Dangora boy, for instance, told me that the Kathariya Tharus were the most backwards and least civilised, because they hardly ever went to school. This emulation of Pahaari values I see as a clear indication of the Tharus' attempts to assimilate the practices and values of the twice-born castes. Let me now describe in more detail how Pahaari customs had become something the Tharus could measure themselves against.

When a Dangora in Geti explained their wedding practices to me, he said that Dangoras earlier used to "sell their daughters" (cf. bride-price). Now this "bad practice" (naramro calan) was abandoned, and Dangoras did like the Pahaaris and payed dowry, where the groom's family gives money or sweets to the bride's family.⁽⁷⁹⁾ The Ranas, he emphasised, would still "sell their daughters". Some Dangora men also made jokes about the Rana female dress. The reason why the Rana women wear big anklets (paela) was to prevent them from running away from their husbands. With two big anklets they would not be able to walk fast or far. Implicit in this argument they gave the Rana women a particular social attribute as unfaithful wives. Although Tharu women can divorce and marry more than once, something high-caste women are generally not allowed to do, this is not considered a good thing to do. A common ideal for a Hindu woman and wife is Sita, the devoted and faithful wife of Rama in the Ramayana epic. Although this ideal may be more prominent among high-caste women who treat their husband "like a god", it is considered bad to leave one's husband both by Ranas and Dangoras. Tharu women, who for various reasons want to leave their husband, must first find another man, which is not always easy.⁽⁸⁰⁾

Another statement Dangoras made about Ranas, was related to their educational standard. Ranas do not send their children to school, they often pointed out, whereas even poor Dangoras send their children to school nowadays. Questions of rank were also discussed. Dangoras often claimed that the Ranas were a lower jaat (tallo jaat) than themselves. Dangoras claimed, for instance, that Ranas originally were leatherworkers (chamaar), who are untouchable and belong to one of the lowest in rank among the occupational groups. Even today Ranas were dirty and hardly ever washed. The same was more or less stated by the Ranas, who would say that the Dangoras were lower than them in the jaat hierarchy. They were also aware of each other's claims. As one Dangora informant told me: "We say that the Ranas are lower jaat, and the Ranas say that we Dangoras are lower jaat". The Ranas in Geti frequently gave the Dangoras one special attribute, which was not mutually claimed by the Dangoras. This attribute was the knowledge and practice of witchcraft.

Not long after I had moved into Gorya ghar, Phulpatti, one of the daughters-in-law whose son was sick, asked me to accompany her to the female healer (bharri). Since the healer lived in the Rana neighbourhood, at the opposite end of the village, we had to cross most of the village to get there. When we passed some Dangora houses, Phulpatti lowered her voice and asked me if I had been in a Dangora house. When I said that I had visited all the houses in the village, also Dangoras, she told me to be careful with Dangoras because they often practised witchcraft. "We never visit Dangoras", she emphasised, "because some of them are witches (boksiis)".

The belief in female witches is common in Nepal and was also common in Geti. A witch is considered to be able to cause sickness and even death through magic formulas (mantras). Very often misfortune is explained as witchcraft (boksiiko kam), and witchcraft is also often used to explain why people do not recover from an illness, in spite of prescribed treatment. Phulpatti's son had, for example, not recovered from his diarrhoea, and the female healer told us that this was because of witchcraft.

The only proper way to deal with witches and their malevolent work is to have the healer identify and undo the spell cast by the witch. The healer, therefore, said some mantras over the boy before we went back home. When we were gathered around the fire, later in the evening, I wanted to know more about the witches and who they were. The Rana women explained that some of the Dangoras in Geti were witches. While they talked about this, they lowered their voices, as if they did not like talking about it, and as if they were afraid others might hear what they said. Except for the healer, nobody knew the identity of the witches.

Women are usually less in contact with other people outside their own kurma and jaat, and I thought that this idea of Dangora witches was something prevailing among the women. A couple of weeks later when, in addition to me, several kurma members were ill, I had the chance to experience how the men related to this issue. Many in the kurma were sick, and aya was sure that it was due to witchcraft. She had therefore called for two of the village healers, who tried to identify the reasons behind the various illnesses.⁽⁸¹⁾ The cause of my illness was witchcraft, they said, and one of them uttered mantras over some rice grains which he gave to me. I wanted to know why a witch had attacked me, but the healers only smiled. Phulpatti, however, argued that I got ill because I visited Dangora villages. "You go to Junapur and Majgaon and other Dangora places - no wonder why you get ill", she said. This raised a debate about Dangoras and witchcraft, and there was a general agreement among all those present that Dangoras and no other jaat practise witchcraft. When I wanted to know how they could be so sure, the answer was: "The healer knows." One of the young boys said that everybody knew that Dangoras practiced witchcraft, because of all the spirits (bhuuts) they kept outside their houses. The spirits he referred to were the clay horses which belong to the Dangoras' pantheon of deities.⁽⁸²⁾

During our discussion of witches they wanted to know whether there were witches in my country (des). As far as I knew, I said, there were no witches in my country, and most people do not believe in the existence of witches nowadays either. "Are there Dangoras in your country?", one of the young men asked. And after I said that I did not know of any Dangoras in Norway, the same person concluded: "How can there be witches if there are no Dangoras?"

This witch topic was raised several times during my stay. Each time I was sick, aya was very angry with me for "roaming too much around" and for drinking beer in Dangora houses. Her youngest son often consumed beer in Dangora houses, without getting sick, I emphasised. "Some Dangoras practice witchcraft, while others do not", she said. And since one could not know who among the Dangoras actually practiced witchcraft, one ought to be careful.

The interesting point here is not that people believe in witches and the rationale behind their beliefs (cf. Evans-Pritchard about Azande). What is interesting is that these witch accusations were only directed against the Dangoras. The Dangoras also believe in witches and spirits, but they never came with specific accusations against the Ranas. When I talked to Dangoras about witches they usually said that witches could be among all the jaats. When I told some of my best Dangora informants that the Ranas accused them for being witches, my informants said that it was typical for the Ranas to do so. Another Dangora informant told me that "the Ranas accuse us of practising witchcraft, and we accuse them of the same", but I never encountered such witch accusations directed especially against Ranas. Witch accusations were most commonly expressed among the Ranas, and the Ranas in Geti were the only ones who gave such a social attribute to a particular jaat. Many Pahaaris as well as some of the Dangoras would distance themselves from the whole "witch thing" and consider it "superstition". They would say that we earlier believed in witches, but not anymore.

Any detailed analysis of ethnicity must take into account the varying cultural significance of ethnicity, not only cross-culturally, but also intra-culturally, and perhaps most important, intra-personally (cf. Eriksen 1992). This was also the case with the Ranas' witch accusations. Aya, for instance, told me that it was all right for me to eat in Sushila's natal home. Although Sushila was a Dangora, aya knew her, and she pointed out that neither Sushila nor any of her family members practised witchcraft. As it seems to me, witchcraft accusations were not solely based on ethnic or caste criteria, and one would perhaps not attribute such characteristics to persons one knew personally.

Some cultural and social attributes are often used and also become important as symbols of identity. Some of these attributes, which are not readily apparent, nonetheless become important as symbols of identity and foci of self-esteem. The mutual "negative Pahaari values" of the Ranas and Dangoras in Geti could be seen as a way for themselves to increase their own status. Dietary preferences, ritual celebrations and ceremonies are also important symbols of identity. I will now turn to contexts where symbols of identity are communicated through dietary habits and marriage relations.

In the introduction, I mentioned the reason why I had established two Tharu homes in Geti. The ward chairman Basu Dev Chaudhary had told me that food was not available in the Rana house I moved into (Gorya ghar ) and that I had to find a place elsewhere to eat. I had therefore arranged to eat in Sushila's natal home which became my Dangora home in the village.

When I left my Rana home at meal times, the women always asked me where I was going. When I told them that I was going to eat in Sushila's house, they seemed surprised. After a while aya even looked angry when I said I was going to eat, and since I did not want to offend neither her nor anyone else in the house, I tried to sneak out of the house. This lasted for a couple of weeks, until the head of the household asked me why I did not want to eat in their house. I was really surprised by this question and told him that I thought food was not available for me. "We are a big family", he said, "and of course we are able to feed an extra person!" He also told me that his mother worried about village gossip if people heard that their big family did not offer me food. After I came to know this, I started to eat in Gorya ghar as well as in Sushila's maiti.

Why had Basu Dev told me that I could not eat in my Rana home? According to Basu Dev himself, he did not think Rana food would be suitable for me. Ranas usually eat "dry rice" (sukkhaa bhaat), he explained, rice with vegetables only, without the lentil soup (dal) or any other kind of sauce (jhol). Dangora food, he pointed out, would always come with dal or jhol. Basu Dev had offered me to eat in his house and being a vegetarian, he reckoned that his food would be more appropriate for me.⁽⁸³⁾

When I had told him that I would eat in Sushila's house, he had nothing against that. Since Sushila is a Dangora, I guess he considered this food appropriate for me. The difference between sukkhaa bhaat and rice which comes with jhol was obviously of importance among the Ranas and Dangoras. Some of the Rana children in Gorya ghar, for instance, told me that they did not like Dangora food because their food came with too much jhol. And many others in Geti would also make such a distinction between their dietary practices.

Food is an important identity symbol in South Asian society - as elsewhere - and is related to issues of purity and impurity as well as to issues of status and social mobility. Among Tharus, all these issues were considered in their discourse and practice around dietary habits. Food symbolises identity at the same time as it marks boundaries. The rules originally laid down in the Muluki Ain were very detailed when it comes to regulations of food and contact between jaats. There were restrictions with regard to whom one could accept food and water from, and food was also hierarchically ranked (see Höfer 1979; Dumont 1980). Vegetarianism and teetotalism were the superior forms of diet, which gave a meat diet and the consumption of alcohol an inferior status.

The boundaries drawn by the discourse and practice around food deal with questions of status. In the caste system, the hierarchy of foodstuffs attributes to each stratum's diet its hierarchical value.⁽⁸⁴⁾ A change in dietary practices has frequently been described as a means for the lower strata to acquire a higher social status. Although it was originally only the Brahmins who became vegetarians (not the Ksatriyas), Tharus in Geti tried to optimise their social status claiming that they were originally Thakurs (Kshatriya varna), and that they had lost their status by consuming meat and alcohol (cf. Chapter 2).

Another commonly described way to improve status, has been to stop accepting water or food from people next to or lower than oneself in status. Many Tharus, for instance, told me that they would not accept everyday food from the other Tharu group. The women in Gorya ghar emphasised strongly that they never had tasted and never would eat Dangora food. This was, aya explained, because of the Dangoras' dietary habits. Dangoras eat buffalo meat (bhaisii ko maasu), which is banned for Hindus in general. Dangoras also eat rats (musa)⁽⁸⁵⁾ and frogs, aya said, food which is considered dirty (phohor) by Ranas.

Other Tharus in Geti emphasised that nowadays, Ranas and Dangoras accept food from each other. The regulations and distinctions based on dietary practices belong to the past. Women seemed more concerned with food regulations than men. Rana men often ate and drank in Dangora houses, and they also ate rats. The contradictions and paradoxes inherent in what people say they do and what they actually do, is here as true as anywhere else. My main Dangora informant would, for instance, eat rice with me in Gorya ghar, and one of my Rana informants would eat rice in Dangora houses. Aya and the two daughters-in-law in Gorya ghar shared their water-pipe (hookah) with the wife of their Dangora kamaiya, and it has been reported as rare to share the water-pipe with others than members of one's own jaat.⁽⁸⁶⁾

Tharus referred to the occupational castes as the lowest caste (sabbhanda tallo). The women of these occupational castes participated in the Tharu women's celebration of carai, where sharing of food is an important part of the celebration. The women of the occupational castes did not bring their food, but they were present and received food from the Tharu women. This is in accordance with the usual caste rules and food restrictions: One can accept food from a caste higher in rank but not from a caste lower than one's own in the local caste hierarchy. There were also several examples of intimate relations between the members of these occupational castes and Tharus. The goldsmith, for instance, had stayed in Gorya ghar and his family would come to visit. The goldsmith himself had made a ritual friendship (mit) with one of the Gorya sons.

To accept food from someone, Dumont argues, can also be seen as acceptance of intermarriage. This is because intermarriages in general presuppose that the two parties can eat together (cf. Dumont 1980). When the Tharus talked about wedding celebrations and that they were going to a wedding, they would normally express metaphorically that they were going to "eat in a wedding" (bihaa khanen).

A general characteristic of caste is that one has to marry within the group.⁽⁸⁷⁾ The various Tharu groups, I have earlier mentioned, have been described as endogamous. The whole idea of endogamy has been criticised, however, and Dumont pointed out that "endogamy is a corollary of hierarchy, rather than a primary principle" (Dumont 1980:113). Endogamy in Nepal, however, has had a special position and functioned as the prime boundary between jaats (cf. Guneratne 1994). What about intercaste marriages in Geti - did they ever occur, and how did people talk about them? When they discussed marriages, Ranas and Dangoras in Geti argued that they kept strictly to their endogamous ideals. "We do not practice intercaste marriage", both emphasised.

In Chapter 2, I briefly mentioned the marriage alliance among the Rana Tharus, which is usually settled during a betrothal ceremony (magni). This ceremony is uniquely a Rana-Rana affair and usually established when the children are about three to five years old. The negotiations often go through a go-between called majpatiya who has been described as some kind of specialist in matchmaking with connections in many villages (see Srivastava 1958. See also Kittelsen 1996). According to my material from Geti, it seemed that the go-between often was a relative or affine of the family looking for a betrothal relation. Often, the go-between would find a person in his own village. Marriage alliances therefore tended to be village-based alliances. The go-between had to make sure that the two parts were of different patrilines (kurma). The use of a go-between who knows both parts was considered important, because there was less risk involved when a third part knew the families. This emphasis on making good and stable marriage alliances seems to coincide with what McDonaugh pointed out: Because the patrilineage is so vaguely defined, marriage alliances become more important.

The negotiated marriage alliances never exceed the Rana community, but this ideal of endogamy does not mean that intercaste marriages never take place. Intercaste marriages do occur but they are not arranged. "Love marriages", known as udhari-udhari, happen quite often. In these cases the young couple run away, and when they come back after a few days their marriage is usually approved/acknowledged. But most love marriages are also within the traditional endogamous units, and seldom intercaste.

In Geti, there was only one mixed Rana-Dangora marriage, and no mixed Tharu-Pahaari marriage. (Sushila had married a Brahmin colleague at the hospital but lived in her husband's village - not in Geti). Narayan Rana had first been married to his fiancée, who later left him. His next Rana wife also left him, and Narayan then married a Dangora girl. Narayan and his family lived in a separate house in another part of the village than his father and brother. When Narayan married the Dangora girl, his father was very angry with him, and Narayan therefore moved to a separate house. His Dangora wife was integrated into his kurma through a ghee-giving ceremony which is common for all new brides in the Rana system. As far as my material indicates, Narayan's family was not socially stigmatised for this marriage outside the traditional jaat boundary. Narayan was not talked about as the Rana who married a Dangora, but a Rana girl from a village near Dhangadhi bazaar was, by the villagers in Geti, known as "the one who ran away with a Chaudhary".

When intercaste marriages occur among the Tharus, there does not seem to be the same restrictions as among some of the twice-born groups.⁽⁸⁸⁾ Intercaste marriages are, however, not encouraged, as Narayan's case and the following example will show.

Amit Chaudhary is a young Dangora from Tikapur in Kailali. His family is wealthy, and Amit has been to very good schools and is now studying journalism in Kathmandu.⁽⁸⁹⁾ When Sarasvati, one of my local Rana informants, started her studies in Kathmandu, Amit met her and liked her very much. Like most Rana girls, Sarasvati already had a fiancé in India, but since he was uneducated, she did not want to marry him. One of Amit's relatives visited Sarasvati's family and proposed a marriage between the two, but Sarasvati's father reacted in a very negative manner. He phoned directly his daughter in Kathmandu and told her that "we do not practise intercaste marriage". Not long after this event Sarasvati's magni relation was broken, and she married another Rana a couple of months later. When it comes to marriages, the traditional jaat boundaries seem to be strong among Tharus in the far west. The ideal marriage relations, both arranged ones as well as "love marriages" were, in Geti and the surrounding villages, in accordance with jaat endogamy and would not go beyond these boundaries.

These examples from Geti indicate that symbols of identity used by the Tharus often communicate social status. In order to enhance their social status, Dangoras and Ranas attributed to each other what I called "negative Pahaari values". This maximisation of social status was based on Pahaari and mainstream Nepalese values, which are similar to the various criteria given the caste system (cf. Introduction). In the contexts above, both Ranas and Dangoras hold on to their mutual distinctiveness as different castes, and the strategies they have chosen, support the caste structure.

There were, however, contexts where Ranas and Dangoras in Geti would express a collective Tharu identity, and this occurred in relation with the Pahaaris.

Expressions of a Tharu "Ethnic" Unity

During my search for a room in Geti, I had especially noticed three large brick houses. The biggest one, a nice red- and green-painted house, belonged to a Chhetri originally from Dadeldhura. The second one, a yellow house with a nice garden, belonged to a Brahmin also from Dadeldhura. He had been a ranger and was commonly referred to as falester. The third one belonged to a Brahmin from Dhoti. He was called Sahukar, a title given to powerful people (thulo manchhe) and also a common term for moneylenders..

This was just before the big celebration of Dashain and people were busy in the fields. There were few people at home, and I was only offered a place to stay in two Rana houses. I finally decided to move into the house where we had been offered to sit down on a rope bed, which is a common thing to do when visitors come.

Sushila commented upon what she considered the Ranas' lack of friendliness. The Ranas of Geti had changed, she said. When she lived in the village, they had always been friendly and invited people to sit down. There were hardly any Ranas left in Geti either, she added. Geti used to be a Rana village, but after the Pahaaris had moved into the village, many of the Ranas left for Kanchanpur and India. This was not only the case in Geti, she said, but in her village Jalari too. In Jalari, nowadays, the Ranas had decided not to sell land to Pahaaris. According to them Pahaaris meant problems (badmaas).

Another thing which had struck me when I was looking for a village, and which I also often experienced later, was that Tharus often said that there were no Pahaaris in their village. When I later came to the village, it turned out to be several Pahaaris settled there. "They do not live in the main village", the Tharus sometimes pointed out, but in many cases the Pahaaris would actually live in the main village. When a person speaks about his village, it is quite common to refer to his "caste fellows" (cf. Dumont and Pocock 1957:26).

A Tharu dominance was visible in the administrative structure of Geti. Although social interaction in general was freer and more intimate within the Tharu population than between Tharus and the twice-born Pahaaris, boundaries were drawn within the Tharu population, and Tharus considered themselves to be people of different origin. But at times they united.

I have already pointed out how identities were locally tied to the village, the patriline and the household.⁽⁹⁰⁾ Most of the Tharu villagers in Geti, emphasised that they had always lived in this area (des). A topic which seemed to preoccupy them and in which they could all participate was tellings of what I here will call the "village past". A village may have as many pasts as it has inhabitants, Cohn has pointed out (1987b). What struck me in Geti, however, was the similarity in these tellings, and I will therefore call these tellings "collective representations" (cf. Malkki 1995). These tellings were always about how Pahaaris had cheated the Tharus.

Collective stories flourish where they have a meaningful, signifying use in the present (cf. Malkki 1995:241). The way in which the villagers in Geti expressed their own vulnerability in relation to the Pahaaris, was told so many times by different persons that it seemed like an institutionalised and commonly known narrative. Let me turn to this collective Tharu representation and describe the message communicated there.

The father of Sahukar was the first Pahaari to settle permanently in Geti. He came from Dhoti, and until a few years ago, Sahukar and his family lived in a mud house. Where Sahukar today has his new brick house, there used to be a big Dangora house. Sahukar asked them to exchange land with him. And this exchange, he argued, would bring the Dangoras closer to their own jaat and himself closer to his jaat. Everybody knew that the soil on the land he offered in exchange was very loose, and that it would be difficult to build a house there. The Dangoras, however, were afraid of not doing what Sahukar wanted, and they also thought it would be nice to live near their own jaat. After this change of land, the Dangora house split into many separate houses which one by one sold their land to Sahukar and left the village. One person left Geti without selling, but he came back after a few years and sold his land to Sahukar. He is now working as a bonded labourer in Attariya, north-west of Geti.

The next Pahaari who came to Geti, was the ranger. He was poor and landless upon arrival and lived in a small house where the village mender lives today. If somebody from the village cut a tree in the jungle, the ranger threatened him with jail. If the person in question paid him money, on the other hand, nothing would happen. Everybody was afraid and paid the ranger the money he asked for; sometimes 500 rupees, other times 1,000, and sometimes even 1,500. This ranger is now a very rich man and owns at least 30 bigha of land. It was on advice from the ranger that Basu Dev Chaudhary's joint family decided to sell some land to Ram Bahadur Khatry. Because of a bad drought, Basu Dev's family had borrowed money from the moneylender and therefore needed money. The ranger had also pointed out that Ram Bahadur, who had political contacts through the panchayat system, would be an asset to the village. Immediately after Ram Bahadur moved into Geti, the village got electricity. Ram Bahadur bought only 0.8 bigha from Basu Dev's family at that time, but as soon as land was for sale, he would buy more. Most of the land that belonged to the Ranas who left Geti, for instance, is now owned by Ram Bahadur.

Another Pahaari moved into the village and opened a liquor store. Until he came, there had been no such store in the village itself. Many Tharus drank a lot of liquor, and the owner gave credit, which we Tharus signed. Since we Tharus were illiterate, the owner added a few zeros to the credit. In order to pay, people were forced to sell their land.

What can be drawn from this story? The first point I want to make, is on the Tharus' perception of their vulnerability vis-à-vis the Pahaaris. The Pahaaris had benefited from the Tharus' illiteracy and weakness, and this has led to their material and socio-economic deprivation. Pahaaris had not always been rich and dominant, and many had actually been poor upon arrival. Because of the Tharus' "ignorance", the Pahaaris had become rich, while Tharus, who used to own their own land, now are landless workers. This has sharpened the Tharus' view of Pahaaris as exploiters and cheaters and fostered anti-Pahaari sentiments among the Tharu villagers.⁽⁹¹⁾ When I asked how the Tharus could let this happen and why they did not fight back, I was always given the same answers: "We Tharus were a straight and honest people (siida jaat). At that time, there was no unity (samaaj thiena). When someone made trouble for us, we would rather move than fight back. We were uneducated, so we did not know how to read or write." Embedded in their Tharu identity is a feeling of powerlessness in relation to the Pahaaris, and their collective representation of the village past mirrors asymmetrical relations of power. It is in many ways a "heavily moral story" (cf. Malkki 1995:53).

When I discussed village issues with some of the Dangora women, they told me that the Pahaaris had caused a lot of trouble (badmaas) to the Tharus. When one of the Dangora women came to Geti about ten years ago, there were only a few Pahaaris, whereas today the Pahaaris "are all over the place". The Pahaaris already settled in the village bought land for outsiders (bhaira ko manche). Some of the Pahaari landowners never settled in the village, but leased out their land instead. The Tharu villagers had therefore decided not to sell land to Pahaaris. "All Pahaaris are the same", the Dangora women said, whereas Ranas and Chaudharys are like brothers (bhaai-daai jastai).

"When Tharus believe that some culturally defined group - not the rich but the Brahmins - benefit disproportionally or unfairly in the allocation of resources, then ethnicity is likely to come as a mask of confrontation" (Guneratne 1994:395). That the twice-born Pahaaris have benefited "disproportionally" and "unfairly", was something the Tharus in Geti commonly expressed.

"For most people", Appadurai points out, "the past is not a predominating, structuring force in their everyday lives, in the sense of a charter to be acted upon" (quoted in Malkki 1995). I will argue, however, that for the Tharus in Geti, the past is "a charter to be acted upon" in certain contexts. Their experience of being cheated by the Pahaaris is now a "charter" Tharus act upon, and I will discuss how this was manifested during the annual village meeting to appoint a new village headman.

This meeting was held in Basu Dev Chaudhary's house, and it was led by Basu Dev himself. As already mentioned, such meetings are usually attended by the household heads or other male representatives from each house. At this meeting so few houses were represented that a messenger was sent around to remind them about the meeting. Finally, most of the houses were represented, but among them only one Pahaari. The meeting nevertheless started. "They never attend our meetings", the Tharus present emphasised. "They do not respect the system."

The Pahaari, a Brahmin, was among the first to settle in Geti and lives nearby Tharus. The Tharus referred to him by the term baaje (grandfather), which is a common term used for Pahaaris, and it is a way of showing respect. The former headman had finished his year and did not want to continue. Basu Dev therefore asked for someone to take over. The Brahmin suggested that the headman position be altered equally every third year between Pahaari, Dangora and Rana, but the Tharus did not agree. Birbal Rana said that the former headman had been a Rana, so now it was time for a Dangora. Dangoras are in majority too, he added. His suggestion was met with support, and people said: "Dangora is all right." The name of one Dangora was mentioned, and Basu Dev approached him with vermilion powder (sindur) to throw on his head as well as a garland. The person in question refused, ran away from the circle and did not come back until another person's name was suggested: Birbal Rana. Basu Dev tried to put a garland around Birbal's neck, but he ran away as well. The garland was instead put around Resham Rana, and powder was thrown into his hair, and Resham was thereby appointed to the headman position. Later the same afternoon it was celebrated, as tradition required, with liquor in the new headman's house.

In between the liquor drinking we discussed the Brahmin's suggestion, and some of the Tharus explained why they did not want the headman position to be passed on from one jaat to another every third year. "We Tharus do not want to have a Pahaari headman, and we have therefore always appointed a Tharu to this position", they emphasised, and someone added: "We Tharus are like brothers." This use of the term Tharu is in itself significant (cf. Chapter 2).

The Pahaaris, I earlier pointed out, do not participate in the collective village rituals in the village shrine, neither do they pay the annual amount of rice to the healer and the headman/watchman. There are no Pahaaris represented in the ward committees either. According to the Pahaaris in Geti, the village headman was a Tharu custom that they did not practice, and they did not seem very eager to have more influence in the village.

The fact that Tharus in Geti wanted to maintain this position within their Tharu community may indicate that the headman still means something to them. Why would it otherwise be so important to exclude the Pahaaris from this position? One may also wonder why the Tharus needed to "vote" together in order to exclude the Pahaaris from positions the Pahaaris probably did not want to have anyway.

Has the balemansa become a symbol of Tharuness, or a symbolic statement of Tharu unity? Or have their earlier experiences produced this general mistrust vis-à-vis the Pahaaris? It is difficult to say exactly what lies behind this collective Tharu action. Historical processes, however, as well as the contemporary distribution of power are often important factors when it comes to how processes of exclusion and inclusion are produced and reproduced (Eriksen 1992:221). This case may therefore be an example of how earlier experiences now influence Tharu collective action.

This meeting, at least, showed a tendency among the Tharus to group together in a mutual exclusion of the Pahaaris. The way the Tharus made distinctions between themselves and the Pahaaris was based on ethnic principles - the communication of cultural differences. Ethnicity, it seems, came in as the most important criterion for excluding the Pahaaris. Processes of social inclusion and exclusion, however, are ambiguous and context-dependent. Srinivas (1996:28) argues that such processes are not only contextual but also a matter of degree.

The question of membership in Geti is in many ways a question of degree. In the context above, Tharus emphasised the need to exclude the Pahaaris from influential positions in the village. In other contexts, however, Tharus complained about the Pahaaris' lack of participation and interest in village matters. The idea that Pahaaris did not do their communal duties, for instance, was strongly expressed.

A few days after the election the headman came around to tell the villagers about community work the next day. The men were going to repair the main village road, whereas the women should clean the village lanes. The participation turned out to be very low, with only 25-30 persons attending. There were no Pahaaris, and many Tharus failed to show up as well.

While the Tharu women - me included - were doing the very dusty job of sweeping the lanes, they complained about the lack of communal interest shown by the villagers nowadays. They also argued that Pahaari women in particular never attended community work and that they were lazy. "We Tharus have to work", they said. The idea that Pahaari women were lazy, was opposite to what the Tharu women had told me earlier about the poor Pahaari women who worked so much and who had no freedom.

The low participation resulted in another meeting in the house of the new headman. This time the whole village was present, Tharus as well as Pahaaris. Many of the villagers had not met earlier, among them, the newly settled Pahaaris of the Achaam neighbourhood. Although some of these Pahaaris had lived in Geti for at least a couple of years, the ward chairman registered their names for the first time.

Basu Dev explained that all villagers had certain duties, and that these duties should be fulfilled by everybody. Sahukar pointed to a problem this compulsory village work caused for people who worked in offices or had jobs outside the village. Most of their children went to school so they could not attend either. An agreement was made, and everyone seemed pleased with this agreement. Those who could not attend, could pay the amount of 40 rupees instead.

During the meeting several Tharus complained about the Pahaaris, who, according to the Tharus, never did their duties. Pahaaris also used the village roads, and they should therefore also help to maintain them. The Pahaaris said that they were never told about such community work. One of the Ranas pointed to the fact that Pahaaris did not give the annual amount of rice to the headman. When one of the Pahaaris argued that the headman was a Tharu custom, Tharus emphasised that the headman looked after the whole village, while the healer (bharra) was a particular Tharu instititution. Basu Dev explained the system of watchman/headman in Geti, and that all the houses ought to pay 15 kg of unhusked rice (dhan) or 60 rupees to the watchman for his job. To show that he agreed, Sahukar sent immediately one of his kamaiyas to bring the 15 kg of rice which he gave to the watchman cum headman there and then.

The villagers also made a decision to implement social sanctions against those who failed to do their duties, such as a boycott of weddings and funerals. The next community work which took place a few days later, was met with greater response by all the villagers. Sahukar, who was not able to attend, had paid 40 rupees instead.

Sahukar had earlier been pointed out to me by several Tharus as a person who did his village duties. He was, my Tharu informants told me, a person who had lived in the village for a long time. From the villagers' point of view, length of stay seemed to be crucial in determining who belonged and who did not belong in the village. The Pahaaris in the Achaam neighbourhood were newly settled people whom the Tharus did not know, and who did not know the village routines.

Ethnicity is the communication of cultural difference, but it is necessary to consider the types of context where ethnic distinctions are made relevant (Eriksen 1992:171). In this chapter I have described contexts where the Ranas and Dangoras communicated their mutual differences. In other contexts, such as the village meeting, we saw that a Tharu unity was expressed vis-à-vis the Pahaaris. This meeting also showed how ambiguous the principles of exclusion and inclusion are, and that we are not, as Eriksen (ibid.) points out, dealing with fixed and unambiguous symbolic systems. In relation to village matters and politics, Ranas and Dangoras grouped together around a common cause, which in this context transgressed the patterns of mutual differentiation described earlier. This may indicate, as many scholars have pointed out, that ethnic oppositions are segmentary in character (see Moerman 1965; Despres 1975a; Eriksen 1992).

Membership does not neccesarily have to do with ethnicity, and as Srinivas (1996) has pointed out: exclusion and inclusion operate at all levels of a caste society. As a general rule, the Pahaaris in Geti were actively kept outside the political life of the village community. The appointment of a village headman showed that Tharus wanted to keep this position to themselves. Nobody wanted to be a headman, though, and the headman has not much of his original power left. The exclusion of Pahaaris in certain contexts, however, cannot be interpreted as evidence of their not being part of the village. While trying to keep the Pahaaris outside, the Tharus at the same time wanted to include them. Their main critique raised against the Pahaaris, however, was that they did not do their collective duties. Tharus wanted the Pahaaris to do their share, such as to give the yearly amount of rice to the watchman. The Pahaaris, who earlier had not been integrated into village life, were also now interested in doing their share for the "common good". What the reason behind this interst may be, is difficult to say. Perhaps they were afraid of a possible Tharu unity against them, or maybe they saw integration as an opening to village participation and influence.

A separate Tharu ethnic identity is not a commonly used criterion for identification among the Tharus in Geti. Kin, length of stay in the village as well as gender roles, class membership and economic issues were also often important criteria. When Tharus in certain contexts express a collective Tharu identity, it is important to ask what else is happening? The historical changes and the Tharus' experience of having lost in relation to the Pahaaris can therefore, to a certain extent, explain the opposition expressed against Pahaaris in the case described above.

In the following chapters I will discuss the process of collective identity formation which is taking place among the Tharu communties all over the Tarai. This process of identity formation, which I have called the pan-Tharu movement, is initiated and led by the Tharu elite. There are several actors and organisations involved and various strategies used in this attempt to form a collective Tharu identity. The opposition against Pahaaris is only one of many unifying symbols. In the next chapter I will introduce the main actors of this movement, and the strategies chosen in their endeavour to form a collective Tharu identity.

Part Two: Unity in Diversity

In Part One, I described the diversity which exists within the Tharu population as a whole. I also pointed out that each Tharu group distinguishes itself from other Tharu groups. The two main Tharu groups in Geti, for instance, were much concerned about their mutual differences and looked upon each other as different kinds of people. Despite this diversity, there is at present a unifying movement which I hereafter will call the pan-Tharu movement. The initiators of this movement now find it meaningful to represent the Tharus as an ethnic entity, and try through various activities to form a feeling of unity among a fragmented people who earlier lacked any sense of "ethnic spirit". In this unification movement, the actors also try to create differences between Tharus and their neighbours in the Tarai. Although the various local Tharu groups share many similarities with their neighbours (cf. Chapter 1), the pan-Tharu actors define Tharus as different from the other communities surrounding them. To create similarities where there are few, and differences where there is a great deal of similarity, is not an easy endeavour and it is, as we will shortly see, also full of paradoxes.

This new and "ethnic" Tharu identity is based on two central ideas. One is that all the Tharus are descendants of the same ancestor and thus the same kind of people. The other is that the Tharus are an indigenous people, Adivaasi, who originated in Nepal.

The process of ethnic incorporation as well as claims to an indigenous status are not solely a Tharu phenomenon. Since 1990, it has become much easier to form voluntary organsiations in Nepal. Some of these are based on religious, racial and ethnic principles. There is, for instance, a pan-Mongol fundamentalist movement, the Mongol National Organisation (MNO) which works to unite all the people of "Mongolian racial stock". The MNO considers the so-called Mongols to be the indigenous people of Nepal, whereas the Indo-Aryans are intruders.⁽⁹²⁾ Several Mongol organisations have lately joined in a United National Front (UNF) which wants to divide Nepal into federal states based on ethnic identity.⁽⁹³⁾ At the same time, the pan-Hindu/pan-Aryan movements of the Gangetic plains, such as the Vishva Hindu Parishad (VHP) (The World Hindu Council), have local branches in several parts of Nepal.

The political and social changes which Nepal has undergone over the last seven years have led many communities to express their ethnic identity more firmly than before. Many communities in Nepal now see their traditions, their myths of origin and their languages as evidence that the group to which they belong has always been a well-defined and homogeneous entity (cf. De Sales 1995).⁽⁹⁴⁾ Among the pan-Tharu actors, this is at present manifest in a cognitive recodification of the Tharu history, culture and character.

Although the main actors of the pan-Tharu movement are an educated, modernised Tharu elite, who see the resources inherent in a united Tharu community, this new and "ethnic" form of self-understanding has spread to a wider spectrum of the Tharu population.

The pan-Tharu ideology is not uncontested. Among the two main Tharu groups in the far west, the pan-Tharu message has so far gained very little support. In their attempt to adjust to a changed environment, the Tharus here have chosen a strategy which is closely linked to the subcontinent and commonly termed "caste-climbing" (cf. Chapter 4). In either of these strategies, however, the past and questions related to historical origin are important. I will in Chapter 6 look closer into two myths of origin that stand in contrast; but each sustains the strategies chosen by the actors themselves.

Chapter 5 The Process of Ethnic Incorporation

Tharu Entrepreneurial Activity

"Political organizations and voluntary associations do not arise spontaneously to reflect the demands of a 'natural' ethnic group. They often precede the existence of a widespread sense of group identity and they play a critical role in shaping it or failing to shape it" (Brass 1974:38).

There are at present several organisations and associations initiated by Tharus. Although a few of these organisations existed before 1990, most of them have been established in the wake of democracy. Until recently, most of them worked solely on a local level, with issues related to the particular group in question. These local organisations were often initiated by local elites who were quicker than others to perceive the resources inherent in organisational activity.

Some of the initiators to these various organisations have recently started a closer cooperation, and the result is today manifest in a pan-Tharu movement. This movement has been initiated mainly for political purposes. By bringing Tharu issues into the political arena, the pan-Tharu actors hope to gain political support from their Tharu brothers. I find it useful to conceptualise them as political entrepreneurs - persons who, in order to obtain power positions, establish themselves as a team (cf. Bailey 1970; Eidheim 1971a).

These Tharu entrepreneurs have realised that there may be a lot to gain in representing a numerically strong community and try - through various activities - to mobilise and recruit clients. Several of the local organisations have been recruited and form the pan-Tharu team. Since clients will give support to an entrepreneur only as long as they benefit from his enterprise, the Tharu entrepreneurs have chosen to bring Tharu issues into politics and to represent themselves as speaking on behalf of the Tharus as a whole.

I shall below discuss some of the various organisations that are active in the pan-Tharu movement. I will start with the oldest and most powerful organisation; the Tharu Welfare Society.

The Tharu Welfare Society

The Tharu Welfare Society (TWS) (Tharu Kalyan Sabhaa) is so far the only pan-Tharu organisation which brings Tharus all over the Tarai together. Guneratne (1994) compares the TWS with Indian caste associations, and he argues that the TWS is "similar to these caste associations in its origin, evolution and activities" (ibid.:324). The caste associations emerged in the nineteenth century, and were in many ways a result of the British preoccupation to "enumerate" and rank every caste group in India (cf. Appadurai 1993; Chakrabarty 1994). Caste associations are composed of members from a single caste, and are often founded by western educated male members (cf. Kolenda 1978). Many of these associations operated originally within the paradigm of "Sanskritisation". With modernisation and economic change they later changed in nature and became important catalysts for political organisation in democratic systems based on competitive politics (Guneratne 1994:323. See also Mitra 1994; Kooiman 1996).

The TWS was founded in 1949 by Keval Chaudhary, a Tharu from Sunsari now commonly referred to as "the father of the Tharus". The criteria for membership in the TWS were both based on ascription (caste) and on achievement (appropriate qualifications). Membership was open to "all Tharus who hold a minimum college degree in any subject either technical or non-technical and are actively engaged in social as well as professionally related activities" (TWS pamphlet). Because of the qualifications needed, the TWS remained an elitist organisation which represented the interests of an elite and its allies on a local level (Guneratne 1994). Wealthy landlords, salaried workers, professionals and college students belong to what Guneratne defines as "elite" (ibid.:321). At the beginning the TWS was not very active and since most of the founding members came from the eastern districts of the Tarai, its activities were restricted to this part (Guneratne 1994). Its core of supporters was thus recruited from among an educated elite mainly based in the eastern part of the Tarai.

The TWS now claims to represent Tharus in every district from Mechi to Mahakali. It has local committees on a district level - also in Kailali, but I never heard of any arrangements or meetings of the Kailali branch of the TWS. The two main Tharu communities in the far west (Ranas and Dangoras) did in general not know about nor relate to the TWS. Although the district committees have their own constitutions, the real power is in the hands of the central "executive" committee based in Kathmandu. This committee consists of 15 members, who are elected by a general assembly for a period of two years.⁽⁹⁵⁾ This committee is made up of a few Tharus who are part of an "urbanised elite", people who have achieved national prominence of some sort. Some of these members have "access to the highest levels of power in Kathmandu" (Guneratne 1994:327). One of the most prominent among this national Tharu elite is Ramanand Prasad Singh. He is originally from Saptari, but lives in Kathmandu. Singh went to school in Dharbhanga, a district in India which shares border with Saptari.⁽⁹⁶⁾ He also studied history at Benares Hindu University, before becoming a lawyer and appointed attorney general by King Mahendra.

Singh has recently been one of the initiators of an ethnic umbrella organisation; The Janjaati Mahasangh, which has become an important political voice for the various ethnic communities in Nepal which now define themselves as indigenous (Janjaatis). The Janjaati Mahasangh is translated as "The Federation of Nationalities". Janjaati is used both in the meaning "indigenous" and "ethnic minority". And in the Hindi-speaking part of North India, Janjaati is commonly used with reference to "tribes". The more well-known term for indigenous in South Asia, Adivaasi, is also commonly used. The terms Janjaati and Adivaasi are used interchangeably by the activists themselves, but not without disagreement.⁽⁹⁷⁾ Singh was one of two delegates sent from this umbrella organisation to the Vienna conference on human rights in 1993.

During a meeting with Singh in his house in Kathmandu, he strongly argued that there is unity in the Tharu diversity. Tharus are an indigenous people with a common origin, he emphasised. Singh, who has travelled all over the Tarai and "studied the Tharus", is convinced about the similarities between the various Tharu languages. In his opinion, the different Tharu groups can understand each other's dialects. The dialects are only derivations of a former ancestral Tharu language, which used to be the lingua franca of the Tarai. Singh also insisted on the "tribal" aspect of the Tharu communities and that Tharus are a samaaj (community) and not a jaat (caste).⁽⁹⁸⁾ Singh pointed to the egalitarian aspects of the Tharu community, a community without internal divisions (such as caste/jaat). The Tharu language, which, according to Singh, has no hierarchical distinction between high and low, male and female, is an indicator of this egalitarian, hence "tribal" aspect. As I already pointed out in chapters 2 and 3, the various Tharu groups actually have internal stratification, both in relation to gender, age and seniority, as well as in relation to hierarchically ranked sub-castes. Tharus are therefore both "caste"-like and "tribe"-like (cf. Guneratne 1994).

The TWS's main activity was devoted to education, but with an ideology totally different from the mass literacy campaign started by BASE (see below). In a national perspective, Tharus were perceived as backwards, and it was thought that education would help raising the status of the group as a whole. (As already pointed out, the normative rules of the caste system adhere to the group and not to the individual. If some caste members succeed in raising their social and ritual status, this is therefore something the whole caste may benefit from.) With its focus on education, the TWS established a hostel for Tharu students in the Tarai border town of Birgunj. This hostel functioned as an important centre of recruitment to the TWS and lately to the pan-Tharu movement.

The early strategy of the Society was much in accordance with what has generally been termed caste-climbing. The objective of caste-climbing is to maximise the kind of credit symbolised in the idiom of the caste system. Although the process of caste-climbing is a rather slow process, the advantage is that it may benefit the whole group (cf. Bailey 1970; Kolenda 1978).

The first TWS convention took place in 1956, and Guneratne has listed the main points of the resolution made during this convention. A large part of the activities was concerned with (i) "Sanskritisation", (ii) social reform and (iii) modernisation. Among the activities categorised as "Sanskritisation" were the campaigns to stop drinking liquor and to make the Tharu women wear blouses and sarees instead of some local dresses. Among the social reforms suggested were the reduction of expenditures on weddings, feasts and dowries, and to lead simple lives like the Marwaris do in order to "make economic progress".⁽⁹⁹⁾ The modernising aspects focused on the need to encourage education and to improve the Tharus' "attention to health" (Guneratne 1994:328).

In 1980, there was a TWS convention in Sarlahi, which was the first meeting where all Tharu groups, including representatives from Nainital and Champaran, were present.⁽¹⁰⁰⁾ After the 1980 convention, Guneratne (1994) argues, the TWS changed its focus and became more involved with modernisation. The TWS realised that the earlier strategy of "Sanskritisation" did not work. Tharus remained weak and on the margins of the system.

The convention in 1980 continued to focus on education but this time with more attention on adult education. By establishing a hostel for Tharu female students, focus was also directed towards women's education. Men should be encouraged to take "an interest in diverse occupations" and to go to towns in order "to aquire modern skills and means of livelihood" (ibid.:329). The TWS also decided to start some kind of cottage industries. In this resolution, there was nothing which could be connected to "Sanskritisation" as it is defined by Srinivas (Guneratne 1994:328).

Since 1980, the need for special quotas for Tharus in education has been one of the Society's demands vis-à-vis the government. This is similar to the system laid down in the Indian constitution which guarantees quotas for "backward/scheduled castes and tribes". The cumulative growth in their process of incorporation is related to the elite's need to "uplift" the whole Tharu community, a community which, they argue, is "backward in every aspect of life".

Although the change of focus took place already in 1980, the instauration of a parliamentarian democracy in 1990 has further enhanced the process of modernisation and made it politically feasible as well as morally powerful to form numerically strong communities. The TWS has seen the need to recruit a greater number of Tharus to the political pan-Tharu movement, and it has therefore started to mobilise and recruit Tharus on all levels.

In April 1995 I attended the fifteenth convention of the TWS, which was arranged near Bhurigaon bazaar in Bardiya district. The convention gave me an apt context to observe the elite's rhetorical discourse which is manifest in their mobilisation, as well as the responses and reactions to it.

The need to reach out to a wider public was an important point made during this convention and it was argued that Tharus in the west needed to be mobilised and "taught". A new resolution was made as well, deciding that the TWS should register as a non-governmental organisation (NGO). With its base in every district, the TWS could be an important facilitator for foreign aid agencies.⁽¹⁰¹⁾ Dilli Bahadur Chaudhary, the chairman of BASE which has attracted a lot of foreign donor agencies (see below), was asked for advice. The committee also asked me, as a foreigner and representative of "international opinion", what kind of perspectives donor agencies prefer. The TWS would focus on the Tharus' status as an indigenous people, and in relation to this there was a discussion about what should be the right term for "indigenous": Janjaati or Adivaasi (see above). The term Janjaati was abandoned in favour of Adivaasi, because Janjaati, one of the delegates pointed out, gives associations to something which is mixed - for instance a mix of several different jaats.

It was also pointed out that Tharus in the backward areas needed to be mobilised and made more conscious. The speeches held at the convention were much like educational lectures. And like the 1995 convention, which took place in one of the more backward areas of the Tarai (pachi sareko tau), the next one would also take place in the far west and backward part. The chairman of the Kailali TWS committee - a Dangora Tharu from Pakalpur village near Geti - welcomed the delegates to the next convention in Kailali.

The local high school in Bhurigaon functioned as dormitory for the delegates, who were placed in classrooms according to the districts they came from. There was also one separate room for the female delegates, who were very few in numbers compared to men. All delegates received a yellow name button to wear on the chest, which made it easy to distinguish the delegates from "ordinary Tharus". On the football field, there was a stage adorned with flowers, where the central committee of the TWS as well as the prime minister, representatives from political parties and other invited guests were seated. The audience sat under a plastic cover facing the scene. This audience consisted mostly of TWS members from the various districts, but also journalists from local and national newspapers had come down to cover the event. The local Tharus were watching from some distance, outside the fences of the football field.

I mentioned the prevailing feeling among Tharus that they, as a group, have lost in the competition over land. This feeling was strongly emphasised in several of the speeches during the Convention. One of the board members of the TWS central committee put it this way: "The population pressure from both sides (the hills and India) is dangerous for the existence of the Tharus. It is the responsibility of the government to protect and preserve the land, property, religion, culture and language."⁽¹⁰²⁾ Both the prime minister and the minister of health and labour were present in Bardiya, where Tharu speakers pointed to their need for special treatment:

Tharus have been neglected, exploited and marginalised, so that we have not reached national mainstream. We have no access to the decision making body either. In our large population of 2.5 million, there is only a few medical doctors and engineers. Only one Tharu is in the first class rank of the civil service of the government. Not a single Tharu is Chief District Officer in any of the 75 districts, nor any District Judge. Tharus never got love, affection or protection from the state. The government should follow the policy of reservation for the Tharus in government services.

Similar claims for special treatment in order to "reach national mainstream" as well as to get better access to positions in the state and public service, have been forwarded by the TWS since 1980. But after the political change in 1990, this claim has become more strongly articulated. Kooiman (1996), for instance, points to the way democracy in India was perceived as a system of government which assigned permanent power to the group largest in number. Numerical classifications, therefore, provided "interested leaders with an argument to start political action in order to secure a larger share of government benefits" (ibid.:2). These benefits are inherent in the Indian quota system.

There is a general feeling among the Tharu elite that they have not got their equal share in the present political system, and many express a feeling of being discriminated against. "Out of the 500 employed by Nepal Bank Limited in Biratnagar, I am the only Tharu", a man from Sunsari pointed out.

Many Tharus feel that they have become losers in the present democratic system. As the chairman of the TWS pointed out: "After the restoration of democracy, Tharus were also optimistic about a better future in line with human rights." Democracy, however, had not helped the Tharus, the same person argued, because "the facilities given to us during dictatorship, were surprisingly snatched away from us after the restoration of democracy".⁽¹⁰³⁾ Chakrabarty (1994) has pointed out how a community's social and economic progress and backwardness could be determined by measuring their share in the number of graduates, official appointments and parliamentary seats (ibid.:150). In the competition for positions, numbers therefore often come in as "a political asset" and an important evidence to demonstrate a "community's poor achievement or unjustified neglect". (See also Cohn 1987a; Appadurai 1993; Kooiman 1996.)

The prime minister, however, could not promise any special focus on Tharus but promised a focus on the upliftment of all backward communities of Nepal. He also emphasised that there was no reason for Tharus to have an inferiority complex and that they should not think of themselves as poor and backward. It is true, he admitted, that most Tharus never had any chance to go to school and due to exploitation "became backwards".

The numerical strength of the Tharu population has become an important argument for the Tharu leaders in their political claims and stimulated the formation of a collective Tharu identity. Tharus now count their own numbers and draw what they consider to be their own boundaries which differ from government figures.

The Tharu elite's emphasis on numbers must also be related to the under-enumeration of Tharus in the eastern districts of the Tarai. Tharus in some of the eastern districts are virtually non-existent in the Nepali census (cf. Chapter 1). Some Tharus have tried to express their own estimation in a more formal (written) manner. One book published both in English and Nepali by the Tharu Reconciliation Centre (Tharu Milan Kendra) of Saptari, is called "Who is Who?". This book gives the names, educational background and professional title of all the Tharus in Saptari. I received the book from the author himself, and, according to him, the book gives "the total population, voters, VDCs, SLC pass, students, politicians, graduates, and professional positions of Saptari district". In a preface to the English version, the author encourages Tharus in other districts to publish the same kind of books. The "Who is Who" lists 63 VDCs, 166 villages and a population figure of 57,787. Since only Tharu names are mentioned, one may easily assume that there are more than 50,000 Tharus in Saptari, a number which differs enormously from the five mentioned in the official 1971 census (cf. Chapter 1). A Tharu from Dang has estimated the Tharu population to be two million. This number would require that almost the entire Maithili- and Bhojpuri-speaking population are ethnically Tharu, which is not the case ( cf. Guneratne 1994). This figure, however, is taken at face value and used in Tharu rhetorics and in their their own counting. During the TWS Assembly, some of the speakers estimated the Tharu population to be 2.5 million, twice the official number of 1.2 million (see above). Many of the delegates strongly argued that the government numbers were wrong (gallat kura). How is it otherwise possible, they asked ironically, that in just a few years more than one million Tharus died?⁽¹⁰⁴⁾ The larger the group the national Tharu elite claims to speak for, the more influence this elite is likely to have at governmental level, especially if it can represent itself as being able to guarantee a bloc vote (Guneratne 1994). A parliamentarian system based on competitive elections has therefore added to the importance of "enumerating communities". (The fact that politicians from various political parties attended the general assembly indicates the resource inherent in a potential Tharu bloc vote.)

"The language problem is the 'nationality problem' writ small" (Geertz 1973:242).

Language is the key symbol of identity in the Nepali census. Unless the census begins to enumerate people by ethnic affiliation, it is only by demarcating themselves linguistically from other Maithili-speakers that Tharus in the eastern Tarai districts can make themselves visible as a distinct people (cf. Guneratne 1994). If one is Tharu, one must by definition speak Tharu. When I discussed linguistic issues with some Tharus from this part, they told me that the Maithili language actually had its roots in Tharu.⁽¹⁰⁵⁾ The Maithilis, who were educated, had "improved" (sudhaareko) the Tharu language and made it into a written language with a dictionary. In other words, the Maithilis had turned the Tharu language into their own. Because Tharus were uneducated and unable to do the same, they had therefore "lost" their language to the Maithilis. The chairman of the Tharu Language and Literary Council of Saptari told me that many years ago - all the way from Mechi to Mahakali - there had been a Tharu kingdom (Tharu raajyaa) with a common Tharu language. Because of influences from outside, different dialects had developed. When I asked him if this was written down somewhere, he mentioned the name of a member of the TWS, "the honourable Ramanand Prasad Singh", who had written about Tharu history.

Language is often an important first step for the creation of national identities and in claiming nationhood (cf. Geertz 1973; Gellner 1983; Handler 1988; Eriksen 1992). Language is in many ways a reflection of national essence, Handler argues, and "a manifestation of what we are as a collectivity" (Handler 1988:161). Therefore, language may be central to national culture and create "intense concern about perceived threats to the language" (ibid.:161). Language has now become an important vehicle of Tharu ethnic identity, a symbol of contrast as well as an emblem of meaning.

Many ethnic strategies relate to the choice of ancestral languages in the definition of self, in the controversy over national languages, and in the educational system (Eriksen 1992:93). During the culture conference in Saptari, a Tharu and professor from Champaran (in India), argued that there should be one Tharu language which all Tharus ought to understand. Even though Tharus do not share such a language, language has become something which Tharu actors now feel they have to preserve. The urge to preserve something which does not exist, presents obvious problems. During a meeting arranged by BASE at their office in Geti, linguistic issues were discussed. A Dangora Tharu and member of parliament (from Congress) pointed out that since the Dangora Tharus form the largest Tharu community, their language ought to be the official Tharu language and the one to be launched on Nepal Radio.

The linguistic issue is therefore not without paradoxes and contradictions. Some of the Tharu tongues are mutually unintelligible, and Nepali and Hindi are the languages that next to the local vernacular most Tharus will know and understand.

At the Bardiya meeting, Ramanand Prasad Singh made a speech in his local Saptari dialect. He urged the Tharus to speak and sing in their mother tongues instead of in Nepali. He also argued that Tharus be taught in their Tharu languages at school and that they needed to make a Tharu dictionary and to publish textbooks in Tharu language. Since I did not understand his dialect, I asked some of the Tharus sitting next to me what Singh was talking about. They were not from the Saptari area, so they did not understand him either. This unintelligibility of the various dialects did not seem to be a problem, because what was crucial here, was the question of identity.

When the only Rana Tharu present asked the audience if he could speak in his mother tongue, the people present answered, "yes, speak your mother tongue". Without realising it, the Rana continued in Nepali, which caused quite a lot of laughter. He changed into his Rana language, a language which many of the eastern Tharus did not understand. Similarly, when the TWS chairman from Dang started his speech in Nepali, he quickly interrupted himself and said: "I have to speak in my own mother tongue." Hearing this, the audience clapped their hands. Another member of the national Tharu elite excused himself for speaking a mixture of his Tharu mother tongue and Nepali: "I have not practised (my Tharu language) for quite a while, so excuse me.

The TWS is today not the only Tharu organisation of national importance. The TWS has recently been challenged by organisations initiated by young Tharus. There are especially two Tharu organisations which have become important actors in the pan-Tharu movement and which now compete for the same resources and support. The organisations initiated by the young Tharu generation have for some time been critical of what they consider an undemocratic, old-fashioned organisation. According to one of the young leaders, the TWS has not managed to do much for the Tharu community in general. He described the TWS members as "people who only talk but who do not know how to work". Since this young person is running his own NGO, he could hardly be seen as a neutral person.

In Bardiya, as well as during earlier conventions (see Guneratne 1994), the younger generation asked for changes within the organisational structure of the Society. The members of the central committee should be elected and not appointed, the younger Tharus pointed out to me. The 1995 convention was to a great extent a confrontation between the young and reform-hungry Tharus and the old leadership. Some of the young Tharus were lobbying extensively, but they did not succeed in their demands. The chairman of the Society explained to me afterwards that the young people wanted quick changes, whereas the members of the central committee wanted to reform more slowly. Despite the disagreements and different ideological platforms, the various organisations are, as we will shortly see, using the same arguments and strategies.

BASE and Tharu Education for Transformation

In Chapter 1, I mentioned a literacy campaign, "Tharu Education for Transformation", which started among the Tharus in Dang. This campaign - later renamed Backwards Society Education (BASE) - defines itself as follows: "Tharu Education for Transformation is a community development project. The overall objective is through a comprehensive literacy programme to empower the Tharu communities for development" (BASE 1993/94:2). BASE evolved around its project director, Dilli Bahadur Chaudhary, and is to a great extent a result of his experiences. BASE is a Dangora based organisation and this reflects its organisational set-up. Through networks based on kinship, BASE has spread to the Tharu communities in the neigbouring districts of the far west (Banke, Bardiya, Kailali and Kanchanpur) and expanded its activities to other areas. The non-formal education,⁽¹⁰⁶⁾ however, still forms the core of its activity.

When introducing the BASE movement and how it all started, I will start with a short description of Dilli's "life history".

Dilli's grandfather was a pretty well-off farmer in Tulsipur, Dang. By putting his thumbprint on a paper he did not understand, the family lost all its land to a local Brahmin.⁽¹⁰⁷⁾ With this loss, the family had to work as kamaiyas (see Chapter 3) on the Brahmin's land. According to Dilli himself, this happened because his grandfather was illiterate. Dilli's father, who tried to make the paper illegal, brought the case to court without success, so Dilli, realising that his society suffered a lot, decided to fight against this. "I was fighting, and I did not keep quiet. The teachers at school told me that, as a Tharu, I had nothing to do there.⁽¹⁰⁸⁾ I went to eight different schools because of fighting with my teachers."

In 1985, five young boys - Dilli, Defu, Diplal, Birbal and Janak - decided to form a development organisation for their community, and the organisation started with 34 members and a literacy programme for the uneducated villagers. Later, through funds from the NGO Non-Frills, they were able to form a 4 H club, running a vegetable, fruit and cash crops development programme (VFC) in 1986. Despite a lot of harassment from local politicians and landlords, the organisation expanded from 34 to 350 members within three years (Cox 1995:14). Dilli was also able to go to Thailand to a tribal rights conference in 1988, through the help of the Non-Frill's director. There he met representatives for the "Asian Adivaasi Committee" , which donated money to his organisation.⁽¹⁰⁹⁾ The Chief District Officer (CDO) in Dang made pressure upon Dilli to stop his activities, and Dilli was twice sent to jail under the Public Security Act (a panchayat law, which made it possible for officials to jail, without a trial, anybody considered a threat to national security (ibid.:14)). "During this period, it was very difficult for me in Dang", Dilli explained:

Since so many of the Tharus from Dang had left and migrated further West (cf. Chapter 1), I spent a lot of time in Tharu villages in the far west. ... I had no money, and I could not afford to go by bus. I walked from village to village and talked to the villagers about how we could improve our situation, and about the importance of education.

After the restoration of democracy in 1990, repression from government decreased, and a British and a Danish volunteer donated kerosene and notebooks. Together with support from other sources, this has led to a tremendous growth of Dilli's organisation, renamed Backwards Society Education (BASE) (Cox 1994:15). With help from the already mentioned Ramanand Prasad Singh, BASE managed to register as a voluntary organisation.⁽¹¹⁰⁾ After registration, funding from government and development agencies was possible.

BASE's slogan is "first focus on education". But education alone could not solve the problems of the Tharus. Dilli, together with Madame - the Danish advisor who has professionalised the organisation - applied for a 113,000 dollar grant from DANIDA in order to start a general health education programme, income-generating activities and to expand BASE's literacy classes.⁽¹¹¹⁾ The expansion of BASE has in the period following the DANIDA grant been tremendous. BASE has now more than 100,000 members and has become the biggest local NGO in Nepal, and with 742 employed, BASE is the biggest employer in West Nepal. BASE has 18 500 students and runs 730 night classes in 325 villages in the districts of Dang, Banke, Bardiya, Kanchanpur, Kailali and Salyan (BASE, 1993/94:3).

Within six months BASE expanded from 80 to 600 night classes, an expansion that may seem rather unbelievable. Dilli explained how this expansion was possible: the internal structure of BASE is much based upon the traditional structure common among the Dangora Tharus of Dang, with the desbandhya-gurwa as the political and religious leader (cf. Chapter 2). The desbandhya-gurwa had traditionally, together with nine other members, formed an area committee. At the village level, the traditional Dangora village headman (mahaton), together with nine other villagers, made up the village committee. Similarly, BASE's organisational structure consists of different levels of committees. At the lowest level is the village committee, a nine-member committee elected by all the BASE households in the village. Then comes the eleven-member area and district committees - elected by the village committees. The central committee is situated at the central BASE office in Tulsipur (Dang), and this is where the decisions are made when it comes to policy making and priorities. Nine of the members of this committee are elected by the area/district committees, and two members are appointed by the president (Dilli himself). There is a district/area manager responsible for the various activities in the area. "The infrastructure was already there for BASE to build upon. We could use the indigenous system as a point of departure", Dilli emphasised.

The migration of Dangoras from Dang had established a network in the far west which also helped in the smooth and rapid expansion. "Our relatives (naataa) were settled in these districts. They came to know about the non-formal education programmes BASE was running and wanted to start such classes themselves."

In Kailali district, there are two area committees. There is one BASE office just outside Geti village, where the area manager, a Dangora from Dang, is supervising the activities.

Poverty as well as too much work are commonly stated reasons for not sending children to school. But this is not the only reason. Although Gorya ghar (my Rana home) was one of the better-off Tharu houses, none of the children went to school. This was not because they were needed at home. The parents told me that they had tried to send their children to school, but that they refused to go. According to the schoolteachers in Geti, Ranas did not "understand" the importance of education, and the Rana percentage of the pupils at Geti school was rather low. Out of a total of 204 boys, Dangora boys formed a great majority of 144 (70 per cent), whereas the Rana boys numbered only 25 (12 per cent). Among the 115 girls, only 8 Ranas were enrolled (7 per cent) - all in grade one. The Dangoras form a majority of the population in the western Tarai districts (approximately 70 per cent, cf. Chapter 1), but the difference in school enrolment among Ranas and Dangoras in and around Geti is greater than the difference in population (see Chapter 3). When it comes to education all over Nepal, girls are given least priority,⁽¹¹²⁾ and they are also the ones who drop out early, e.g., after the first year. The teachers told me that every year they would go from house to house to persuade the parents to send their children to school. So far, this had had very little effect on the Ranas, whose drop-out rate was very high. Rana children would start, but soon drop out, the teachers told me, and the reason given by their parents was all the work at home. One educated high-caste woman who works at an office in Dhangadhi, told me that Ranas had hardly changed. Their traditions remained "99 per cent the same as many years earlier". Another reason given for the Ranas' seeming reluctance to "modernise", is that they look upon themselves as first and foremost farmers (kisaan). Most of the Ranas own land, whereas the Dangoras, who arrived later, have more often become landless. There is hardly any Rana kamaiyas, and Ranas in the villages I visited emphasised that only Dangoras worked as kamaiyas. The Dangoras, it is often argued, have been forced to seek other opportunities. At the hospital, for instance, the Dangoras form a majority - whereas there is only one Rana employed.

With the slogan "knowledge is power", the main objective of the education activity of BASE is to increase the literacy rate among the female part and the kamaiyas. The non-formal education (NFE) activity is by far the largest, but formal education activities also play an important part. The formal education activities consist of night classes for children who are not able to join the government day school. They use the government curriculum, and local schools cooperate with BASE, so that the BASE students can attend the government exams. BASE is also coaching classes for students preparing their school-leaving certificate (SLC), as well as a Campus support programme which provides scholarships for college students.

The non-formal education activities are generally divided between an "out-of-school children programme" (OSP), for children between six and fourteen years old, and adult literacy classes (ALC) for those above fourteen years. Children following the OSP classes' nine-months' course, are able to join grade four in the government schools. The ALC classes are divided into three six-months' courses which teach the students basic reading and writing ability, as well as health and human rights issues. After this course, students are introduced to other BASE activities, such as women groups,⁽¹¹³⁾ saving groups and income-generating activities.

The curriculum used is provided by the Ministry of Education and especially prepared for such NFE activities.⁽¹¹⁴⁾ The night-class teachers are generally recruited among the villagers themselves and this will have a positive impact on the enrolment of the Tharus. Very often, according to BASE, Tharus do not speak Nepali and feel shy vis-à-vis the government teachers, who are normally of the twice-born castes. BASE has also its own teachers trainer corps, providing training to the night-class teachers and facilitators. This training, with the curriculum as a point of departure, is more practical than theoretical. The topics to be taught are presented, and after a discussion of the different topics of the books, each of the night-class teachers has to get up and practise in front of the other teachers. BASE demands that these have fulfilled class eight.⁽¹¹⁵⁾

When it comes to the kamaiya education programme, one project has been to collect "kamaiya life stories" in order to make the school books more relevant to the life situations of the kamaiyas. According to BASE statistics, the average drop-out rates from the BASE classes vary between 13 and 15 per cent and are much lower than the drop-out rates in government-run schools. And where government-run schools demand a 33 per cent correct score in order to pass, BASE demands a 50 per cent correct score (BASE 1993/94).

In Geti, there are two NFE night classes, one adult literacy class (ALC) for women and one out-of-school programme (OSP) for boys. Basu Dev Chaudhary, one of the ward chairmen in Geti, is also the chairman of the BASE village committee. According to Basu Dev himself, it was much because of him that the BASE programmes were established in the village. "I came to know about BASE through 'Madame'", he said:

I wanted to know more about BASE, and "Madame" encouraged me to go to Tulsipur. In Tulsipur, they examined me and found that I was able to run such an NFE programme. After I came back to Geti, I talked to other people and invited the village headmen and chairmen of the neighbouring villages. Soon we had classes running in several villages in Kailali. Our relatives in Kanchanpur also heard about BASE. They came to us in Kailali to look and learn, and later established their own programmes in Kanchanpur.

The two ward chairmen of Geti teach one class each. In the ALC class, there were about 20 women/girls gathered every night in Basu Dev's house. This class had a clear Dangora majority. There was only one Rana girl and no Pahaaris. When I asked the Rana girls and women why they did not come to the BASE class, they said that Basu Dev's house was very far from their homes. It was on the other side of the village, and they were afraid to walk home in the dark. When there had been a night-class in a house in the Rana quarter, however, many Rana girls had joined. The Rana girl who actually came to the BASE class, lived very near to Basu Dev's house.

Since this ALC class was almost 100 per cent Dangora, the local Dangora language was used more than Nepali. Many Tharu women in Geti, as I already have pointed out, did not speak Nepali very well. The OSP class was held in a Rana house in the Rana tol, and both Rana and Dangora boys attended. There were no Pahaaris.

The curriculum provided by the Ministry of Education was in Nepali, but the teaching was both in the local Tharu languages and in Nepali. The boys used Nepali more frequently than the women, although the importance of learning Nepali was emphasised in both classes. The curriculum is a national one which is used all over the country. There was thus no specific Tharu reality expressed in the books, which dealt with general problems experienced by poor, illiterate people.

The various topics were national geography and history, the importance of clean water and a general good hygiene, as well as the importance of family planning. Environmental awareness and the problems of deforestation were mentioned, and the bank system was also explained: how to open an account in order to save money, as well as the advantage with bank loans instead of loans from moneylenders. There were drawings illustrating the topics, and these drawings were from the hilly part of rural "standard" Nepal.

Although this literature did not express a specific Tharu reality, the problems discussed were related to the local context. Birbal Rana, the teacher of the OSP class, did not talk about the Tharus' status as uneducated and "backwards" (pachi sareko) as a unique Tharu problem. It was referred to as a farmer's problem (kisaan ko samasya). A farmer has to work so much, and because of that, he is uneducated (ashikshit). In order to improve and change the situation of the farmer, education is necessary. The farmers are poor because they are uneducated, he added. Birbal Rana himself had only been able to finish grade six in a government school.

In December 1994, Dilli was granted the Rebook Human Rights award for his organisation's work against bonded labour, and he went to the United States to receive the award. Dilli has become an international person, which again has had positive effects on the government's attitude towards the problem of bonded labour. When Dilli came back from the USA, there was a big welcome ceremony in Tulsipur. A long line of people queued to give him flower garlands (mallas). This welcome ceremony, of the same nature as the ones given ministers and other "big people", indicates the importance now accorded Dilli and BASE.

In a culture dominated by patron-client structures (cf. Bista 1991), Dilli is now considered - to quote Bista - "a person one may approach when needed". This chakari behaviour, as Bista calls it, makes it necessary to be "close to or in the presence of the person whose favour is desired" (ibid.:5). Dilli is considerd to be a strong leader who has both political capital (actual political resources) and political credit (political resources which the supporters believe he has) (cf. Bailey 1970). One of the local politicians from the Congress Party spends more time at the BASE office than anywhere else, the BASE staff members said.

"BASE has now become so powerful that the politicians know we can change local government immediately", Dilli pointed out to me, adding: "We are, by the way, not interested in doing so." During the Congress government, in 1993, the Home Minister's office asked Dilli to be vice-chairman of the Social Welfare Council, something he refused: "The Vice-Chairman is a political appointment, and if I agree, I have to do party politics." Dilli was also asked to be chairman of the Asian Adivaasi Committee, an organisation he has been in touch with since 1988. "I said no, because I have to devote myself to BASE first and foremost."

At the time of the general elections in November 1994, there was a lot of tension in Dang. Clashes between supporters of the two parties, the CPN-UML (Communist Party) and the NC (Nepali Congress), led to deaths and many people were wounded. To avoid the pressure put on him by the local politicians, as well as the danger involved in being a powerful community leader, Dilli ran away to India for one week.⁽¹¹⁶⁾

McDonaugh (1989), who discusses aspects of the Tharu mythology, argues that a pan-Tharu identity has little chance to develop among the various Tharu groups, unless it becomes related to the problem of land distribution and economic questions. This is exactly what has happened with the BASE movement and its focus on land distribution and the kamaiya problem. The problems of bonded labour and landlessness (sukumbaasi) were first raised by BASE as a problem among the Tharu communities in Dang and in the far western region. The support BASE has received from international donor agencies for its focus on the bonded labour problem, has influenced the government's policy and made it into a national issue. Bonded labour has also become a unifying force in the pan-Tharu movement. It has been embraced by the TWS as a political argument to use against the government in its claim for special treatment of the Tharus and in its attempts to mobilise and recruit Tharus in the far western districts. When the Tharu Welfare Society arranged its convention in 1995, the kamaiya problem had become a main topic. In one of the speeches, the chairman of the TWS argued:

We Tharus have lived here much before the Nepali nation state came into being. We settled in dense forests, we cultivated and made the soil fertile. Today, the owners of this Tarai land are landless and even bonded laborers. We were able to struggle with nature, but we did not know how to fight against the cunning and tricky game of human beings.

Until BASE got registered as an NGO in 1991, there was no Tharu organisation in the western part of Nepal dealing with "Tharu problems". The Tharu Welfare Society (TWS) was, as already pointed out, an elite organisation with little influence among Tharus on a local grassroots level. The present chairman of the TWS, who is a former minister of education (during the panchayat system) was earlier working against Dilli, whom he considered a communist and troublemaker.

Very little formal contact, therefore, existed between the various Tharu organisations in the east and those in the west. There was a general attitude among Tharus in the eastern part, that the Tharus in the west were very "backwards" (cf. Ødegaard 1997b). They were looked down upon and Tharus in the east did not want to be associated with those in the far west. But with its national and international recognition, BASE has become an important factor in shaping contacts between Tharus in the east and the west. Recognition received at higher levels has strengthened BASE and Dilli's position in the Tharu environment. Following the expansion and the international financial support of BASE, Tharu organisations from the central and eastern Tarai now come to Tulsipur, eager to keep in touch with BASE and to learn from the BASE experience.

BASE ows its expansion much due to kinship networks and "moral commitment" to a cause. It has, however, developed into an important organisation with a powerful leader, and both Tharus and non-Tharus, politicians and donor agencies seek relations with Dilli and BASE. These new relations has made BASE, to an increasing degree, into a transactional contract organisation (cf. Bailey 1970), where recruitment is based more on the basis of profit or potential profit than upon an ideology or love for a leader (ibid.).

The need to protect "Tharu culture" has been emphasised in the BASE reports as follows: "The Tharu culture is very old and rich in tradition, though the loss of land has created a crisis of cultural identity. Since 1992, BASE has arranged a culture programme in connection with the Tharu New Year (magh sankranti), in mid-January" (BASE 1993/94:30).

According to BASE, the huge outmigration of Tharus from Dang - in addition to the loss of land and socio-economic degradation - has had severe impact on the Dangora Tharu culture. Among the Tharus who left Dang, there were many who took with them their knowledge of local medicinal plants. Therefore, according to Cox (1994), very few gurwas (shamans) in Dang still have this knowledge of herbal medicine. It is also difficult to find young Tharus interested in learning about medicinal plants. As a result, Cox argues, the local knowledge of herbal medicine among Tharus in Dang is in danger of being lost (ibid.:11). In addition to "lost knowledge" of and lack of interest in herbal medicine, Cox also mentions festivals which are no longer celebrated. This is, according to Cox, because of lack of money (ibid.:11). Cox mentions, for instance, a great summer festival which used to be celebrated for three weeks among the Tharus in Dang. This festival, known as barka naac ("the great dance"), consisted of various group dances, singing performances and rituals. But for over 20 years now, the barka naac has not been celebrated at all.⁽¹¹⁷⁾ Other major festivals have met the same fate, and even the most important of the annually celebrated festivals - the magh sankranti (Tharu new year) which used to be celebrated for nine days - is now only celebrated for four days (ibid.).

Tharus had probably not more money in the "good old days" than they have nowadays, and this makes it difficult to accept BASE's and Cox's claim that lack of money is the reason why such festivals are not celebrated anymore. Since so many Tharus have migrated westwards, lack of people to arrange these big festivals may be a more likely explanation.

These culture programmes attract Tharus from the eastern part of the Tarai as well, and more prominent guests also attend. The crisis of cultural identity, now expressed by BASE, has also fostered an interest in local legends. During one of my visits to Tulsipur, I heard about a BASE project of writing down the Dangora version of the history of the valley of Dang. Dangora Tharus ought to learn about their history, it was argued. According to this legend, the valley of Dang was divided between four brothers into separate ritual territories. Each territory was ruled by a religious and political leader, the desbandhya-gurwa, who Krauskopff described as "the perfect magician".⁽¹¹⁸⁾ In 1995, the culture programme was called "Let Us Protect Tharu Culture" (Tharu Samskrti Bachau Karyakram) and it took place at the Geta hospital just outside Geti village, an event I will come back to in Chapter 6.

BASE is not alone in emphasising the need to protect Tharu culture, however. There is at present also an organisation called Tharu Culture Society, which has as its raison d' être to promote and protect Tharu culture.

The Tharu Culture Society

"...a cultural movement which begins largely as a means of strengthening solidarity so that political gains can be made may become genuinely creative, though it must continue to satisfy the requirements of its original impetus" (Orans 1965:132).

The Tharu Culture Society (Tharu Samskrti Samaaj) is the third pan-Tharu organisation I will mention here. The chairman of the Society is a young Tharu from Deukheri, Dinesh Kumar Chaudhary. Dinesh, who comes from a rich landowning family, has studied law in Kathmandu. The main aim of this organisation is to promote contacts between the various Tharu groups, as well as to preserve and promote Tharu culture.

The Tharu Culture Society has established local committees in various districts, and it was its local branch at Bakdhuwa in Saptari which was in charge of the big Tharu culture conference in 1994. The conference, which was called "The First International Tharu Culture Conference", received financial support from BASE. The conference was attended by Tharus from India and Nepal (all districts except Kanchanpur), as well as national and international guests. During the three days' programme, there were various competitions in "folk culture", such as dance, poetry, songs etc. Tharu dance troupes from all over the Tarai were contesting for the various prizes given. One of the troupes was initiated by Amit, a young Dangora who we already met in Chapter 4. At the time of the conference, he was only 16 years old.

Amit himself participated also individually in different competitions, and he was placed third and fifth in the diciplines known as "folk culture" and "poetry". He was also given the first prize in a dicipline called "the best judgement" during the conference. The winner of the dance contest in Saptari was a group of girls from Deukheri, and they had earlier also been to Germany to perform.

In his welcome speech in Saptari, Dinesh stressed that "we should not say Tharus from east and Tharus from west. We should instead say Tharus of Nepal. We have to unite in order to get strong" (cf. Gurung and Korvald 1994). Therefore, marriage alliances ought to be established across jaat boundaries. Endogamy, I pointed out earlier, is an important boundary marker between the different jaats in Nepal. It has been and to a great extent still is one of the main boundary markers between the various Tharu groups (cf. Chapter 4).

A few years ago, the TWS made a decision to abolish the distinctions of group endogamy. This decision, however, had little practical consequences for the Tharu population at large. The symbolic meaning of this was "to recognize that those entering into the relationship are the same kind of people, belonging to the same social universe" (Guneratne 1994:340).

Several marriages between Tharus in the east and the west were negotiated in Saptari. One of them was the marriage between Dinesh and a girl from Sunsari, Sita-Ba Chaudhary. Dilli, the chairman of BASE, had been one of the negotiators, and BASE would also sponsor the wedding by paying the bus fare needed to bring Dinesh and his followers (barait) to Sunsari. Although the initiators of the culture conference in Saptari also belong to those I have categorised as the Tharu elite, there were quite a lot of "ordinary" Tharus present. The dancers who had come to participate in the dance competition, for example, were recruited from the local community level. This made the "audience" in Saptari more heterogeneous in nature than what it usually is during TWS conventions, dominated as they are by the male Tharu elite. Because of the more heterogeneous participation, the symbolic statement of "ethnic" unity which is inherent in these marriage alliances, may have reached out to a wider audience.

The marriage between Dinesh and Sita-Ba took place in May the same year, and it received a lot of attention in the pan-Tharu environment. When the bus left Dinesh's home in Deukheri for the long journey eastwards, it was only half-full. All along the Tarai belt, however, the bus picked up new people: Tharus from Kathmandu as well as from the other Tarai districts. When Dinesh and his followers (barait) arrived in Sunsari, it included people far beyond Dinesh's relatives. The wedding had become a collective Tharu affair and a symbolic statement of their unity. This wedding also functioned as a "cultural meeting" between Tharu women from Sunsari and Deukheri. (The delegates to Tharu conferences and meetings are to a great extent males. Most of the women attending this wedding had therefore not been to far away parts of the Tarai, nor met Tharus from other and distant places.)

Usually men are the ones who follow the bridegroom to the bride's house. When Dinesh's barait left Deukheri, several women came along as well - both his kanchi-ama⁽¹¹⁹⁾ as well as his sisters. Just before leaving, his sisters put on their traditional dress, which consists of a brightly coloured skirt (lahenga) and a sleeveless blouse (lungi). They also put on big silver necklaces and bracelets. Their mother told me that this was in order to show their "culture" (culture dekhaune ko lagi) to the people in Sunsari. Usually, his sisters would not wear such clothes because they go to school.

The attempt to create cultural similarities between Tharu groups who may not have much in commen, is not unproblematic. The wedding in Sunsari brought together women from different parts of the Tarai who earlier had never met. For many of these women, the wedding was a meeting with different traditions and values. And this was something they pointed out to each other.

Upon our arrival in the bride's house we were invited to sit down, and some women came to wash our feet. This is not a common practice in weddings among the Tharu communities in the far west, and one of the women from Deukheri commented that the customs (calan) here were different. Later, when the women of the bride's family had invited Dinesh's female relatives to eat, the food proved to be a totally different experience for the women from Deukheri. Yoghurt, fried breads (puri) and sweets also came with the meal (thaali). Although it is considered bad to be greedy, the women from Deukheri ate so little that the bride's relatives complained loudly. The women from Deukheri said they were not hungry, but they told me that they did not like puri and yoghurt. After the meal, we had to wait a long time before the wedding ceremony kanyaa daan ("gift of a virgin") started, and this waiting gave a lot of time to socialise. Like most Tharus in the east, the women in Sunsari spoke Nepali fluently, and the women from Deukheri had been to school and spoke Nepali too. The talking was therefore done in Nepali, and when someone forgot and talked in her local vernacular, the others would quickly interrupt and say: "Speak Nepali, we do not understand what you say." After some time, however, they started to ask each other about words and customs: "What do you call this, and how do you do that?", and after discussing this for a while, they found similarities as well as differences.

The different ways of dressing were also something commented upon. The Sunsari women found the Dangora clothes indecent and backwards, and the Dangora dress became a focus of discussion. "Do you really wear these kinds of clothes?", the Sunsari women wondered. Dinesh's sisters pointed out that they had only put them on to "show their culture". "We go to school, and normally we do not wear these clothes", they emphasised. The women from Deukheri, on their side, told me that they did not like the way the Sunsari women were dressed either - because of the tail at the back of their dress (cf. Chapter 1). When the bride's friends and relatives came to Deukheri, the mutual resentment in relation to each other' s dresses seemed to have declined. One afternoon, on the rooftop of Dinesh' s house, the women were photographed together in the various dresses, and some of them even exchanged dresses.

A political entrepreneur is a person who tries to attain power positions by recruiting a team of clients. And an entrepreneur is often quicker than others to see possibilities in a changed environment. The Tharu entrepreneurs desribed here are all trying to adjust to a context which has brought new possibilities for organisational activities. The most powerful Tharu organisation has been contested by other organisations led by younger Tharus. They now compete for the same support (Tharu clients) and the same benefit (political power), and they can in many ways be seen as political contestants.

Despite their mutual competition, they have seen the need for joint action and being united around issues which appeal to the international community. They have chosen to focus on global issues such as indigenous and human rights. The TWS, for instance, has made the problem of bonded labour into one of their main arguments - an issue which originally belonged to BASE.

The Tharu entrepreneurs have also realised that there is a great potential in the development business and hope to make their expertise (accumulated knowledge about a problem and/or area) relevant vis-à-vis the international donor community. During the TWS assembly, a decision was made to register as an NGO. This NGO should function as a facilitator and consultant for the various donors wishing to start development activities among the "backward" Tharu communities. Their emphasis on donor benevolence must be related to BASE and all the support BASE has received from international donors for its fight against bonded labour. Many donors have been queueing to support BASE (cf. Ødegaard 1997b and c), and because of Dilli's recognition abroad, he has been offered several important positions.

In their unification process, the pan-Tharu entrepreneurs try to establish a collective Tharu identity based on a common origin and their status as an indigenous people (Adivaasi) of Nepal. This ideology is not uncontested. I will now describe a local organisation which has recently been founded among the Ranas in Kailali and Kanchanpur. The initiators to this Rana Reform Movement claim that the Ranas are a twice-born caste from India. The Ranas' reform movement is therefore in sharp contrast to the ideas forwarded by the pan-Tharu ideologues. The Ranas of Nepal, as we will shortly see, identify strongly with the Ranas in India, and they have chosen to join in a pan-Rana movement.

The Rana Reform Movement

Despite the tremendous support BASE has received in the far west, it is a Dangora-dominated movement. Some of the Ranas, for instance, argue that BASE is a Dangora organisation, and that the Ranas do not get jobs in BASE.⁽¹²⁰⁾ There is at least one Rana employed at the BASE office in Geti, and in Kanchanpur, many Rana villages have been recruited to BASE. That Ranas have not been recruited as easily as the Dangoras is something the BASE staff also admits. Dilli mentioned several times that he wished to recruit and mobilise the Ranas better.

Ranas have instead formed their own organisation called the "Rana Reform Society" (Rana Samaaj Sudhaar). Some of the issues of this Rana reform movement - to modernise the Ranas and to improve their educational standard - are comparable to the main issues forwarded by the BASE movement.

In November 1993, a big Rana meeting was held in Dhangadhi, where about 250 Ranas from Kailali and Kanchanpur were gathered. Most of the attendants were males, and only a few women were present. None of the Ranas in Geti attended this meeting, and they had not heard about the Rana Reform Society. They knew some of the initiators, however, who belong to a rich, educated Rana elite. The purpose of the meeting in Dhangadhi was to form a constitution. No decisions were made, but suggestions and propositions for a constitution were discussed. In March 1994, another and similar meeting was held in Kanchanpur, and written invitations to this meeting were distributed.

The main issues during these two meetings were how to reform (sudhaarne) the Rana community and how to become more "developed" (agadi banaune). Ranas, it was argued, had to abandon "bad customs" (naramro calan) such as the betrothal ceremony (magni) and, just like the Pahaaris, marry at a later age. Ranas also ought to send their children to school and learn Nepali. Another area of priority was to improve Ranas' health and make them use the doctor instead of the healer (bharra). Ranas, according to one of the initiators to the reform movement, were only concerned about basic things such as to live, to eat and to sleep. Ranas had to change their attitude, he argued, so that they would be able to participate in the Nepalese system on equal terms with the rest of the nation.

The initiators to this Rana Reform Movement are two well-educated Rana men. One of them, Suresh Rana, works at the Dhangadhi Town Hall. The other one, Gopal Raj Rana, comes from a rich landowning family. He is a politically active person and member of the National Democratic Party (NDP). The NDP is known as a reactionary party of ex-Panchas, those who were active during the panchayat system. Also Gopal Raj's father, Hari Lal Rana, was politically active during the panchayat system. Although Hari Lal now has refrained from political activities, he is still one of the most influential Ranas. No representative from the Rana Reform Movement had been present at the culture conference in Saptari. In fact, no Ranas at all attended this conference, and neither of the Rana reformers had heard about it. The reason why they had not received any invitation, was probably because the Rana Samaaj Sudhaar was not formally registered, they added. The Rana reformers agreed that cooperation between the various Tharu organisations was positive, but they also emphasised that the Ranas had to start their own reform movement. Since there are so many Tharu groups differing in several aspects, a collective reform movement becomes difficult. Each group has to start with itself, Hari Lal pointed out.

The Rana Reform Movement has existed since 1991, and they want to register as an NGO in order to receive governmental support. According to Gopal Raj, the Ranas of Nepal had never received anything from the government, and he referred to the Ranas in Nainital, who had received a lot of support from the Indian government. The Ranas in Nainital, therefore, were very "developed" and worked in government service.

Since the Indian independence and the introduction of the quota systems, Tharus in India have been given special treatment. In 1967 the Tharus were, together with five other "tribes" in Uttar Pradesh (UP), declared a "scheduled tribe", and the government of UP has also established a special "tribal" court in Nainital (cf. Tripathi 1982). In 1967, the Rana Tharu Council (Rana Tharu Parishad) was established in Nainital. This Council has worked to improve the Ranas' socio-economical and political situation. In the years between 1969-1974, various youth organisations were established, and in 1976 the Tharu Tribal Development Municipality (Tharu Janjati Vikas Nigam ) was established to undertake development activities among the scheduled tribes of UP (cf. Hasan 1992). The Ranas in Nainital have for some years been politically active, and they have also been able to forward certain claims vis-à-vis the authorities. In 1981, for instance, they organised a demonstration to get back the land they had lost (cf. Tripathi 1982).⁽¹²¹⁾

When it comes to social reform movements, they seem to have a long tradition among the Ranas in Nainital. In Srivastava's monography from 1958, he mentioned a social reform movement initiated by a handful of educated Ranas. This movement passed through two different stages. The aim of the first stage, which started in the 1930s, was to abandon old customs and practices. The second stage, which followed the wake of India's independence in 1947, focused on "aspects of social life". This change of focus took place because the leaders realised that their plans had been too ambitious. They also wanted to take advantage of the new democratic India and the emerging elevation of the lower castes after independence. This, they thought, would make the authorities a better listener to their claim of a higher social status. According to Srivastava, the Rana Tharus in Nainital claimed descent from the Rajputs (Thakur) of Chittogarh in Rajasthan, who were driven out by the Mughal emperor Akbar (see Srivastava 1958:14). The strategy chosen by these reformers was in line with what has commonly been termed caste-climbing and they propagated their ideas in open air lectures. In their attempt to climb the caste-ladder, the reformers introduced 18 rules with accompanying penalties if violating them. These rules were in line with the practices and values of the twice-born castes: The Ranas should wear the holy cord like the other twice-born castes do. Offering of meat and liquor during all ceremonies ought to be banned, and poultry raising as well as the consumption of pork were forbidden. There was a prohibition on brideprice, intercaste marriage and on sharing water pipe with lower castes. The Rana women were not allowed to move around freely, and they should not enter the kitchen during their menstruating periods. The reformers introduced restrictions on commonsality and deemed divorce illegal. Ranas were neither allowed to sell cows or bullocks to Muslims, nor to buy meat from, or have their hair cut, by Muslims.

A sort of consciousness was raised among the Rana population, and fines were collected. But instead of using the money for any constructive purposes, they "served to swell the private purses of the reformers" (Srivastava 1958:106). Because of the reformers' lack of credibility, this first stage did not get much support. Until a revival started in 1948, the movement was at the point of death. The reforms suggested this time were also attempts to emulate the values and practices of the twice-born castes. This time, they were more focused on socio-religious issues, and Srivastava divided them into four different types.

The first type of rules were related to cleanliness. The Rana women ought to put on clean clothes before cooking meals and they were not allowed to enter the kitchen during menstruation. The second type dealt with rules for ceremonies. The Ranas should use Brahmins for the various ceremonies, which ought to be performed according to more orthodox Hindu rites. The third type of rules, which Srivastava related to economic issues, banned the use of liquor during ceremonies. A fine of 50 rupees was also introduced to both parts if an engagement (magni ) was broken. The last type of rules dealt with prestige. The Rana women were not allowed to visit market places, and if a woman was found in the market, her husband would be fined an amount of 25 rupees. The second stage was at first met with great response, but since many of these reforms were in conflict with the material and religious needs of the Ranas, the movement lost its force (Srivastava 1958).

More recently, Hasan (1992) has described various reform movements among the Ranas in Kheri district. These reform movements have had a strong impact on the Ranas' perceptions of their own value system. The tendency to emulate mainstream Hindu customs, rituals and observances, has led to puritanic attitudes and a change in the status of the Rana women. One of the reformist movements which have gained a strong hold among the Ranas, is the Radha Swami sect. This movement, which has its origin in Punjab, proclaims austerity, such as vegetarianism and teetotalism. The Rana women who follow this sect, for instance, have stopped wearing their traditional Rana dress and instead wear sarees. The Ranas in Kheri have also established their Rana Tharu Council (Rana Tharu Parishad). After the Hindu nationalist People's Party of India (Bharatiya Janata Party [BJP]) came into power in Uttar Pradesh in 1991, the Rana Tharu Council has established close contacts with BJP politicians (cf. Hasan 1992). I will come back to this in Chapter 6.

Whereas the pan-Tharu movement has its focus on Nepal, the Ranas of Nepal seem to have joined a pan-Rana movement which crosses the border into India. The Ranas of Nepal, in many ways, identify more strongly with their relatives in India than with the rest of the Nepalese population. But they do not necessarily exclude themselves from the rest of the nation, as the example below will show.

In 1994, the national celebrations of the spring festival Holi took place one month later than usual.⁽¹²²⁾ This was unnoticed by the Ranas in Geti who had made everything ready to start the Holi celebration as they normally do: on the full moon of magh (January-February). When the villagers heard that the rest of Nepal's Holi celebrations would take place one month later, they insisted to celebrate according to their tradition. The village headman (balemansa), himself a Rana, emphasised that Holi would go on as usual. A couple of days later, one of the ward chairmen told me that the whole village would celebrate Holi together after one month. When I came back to the village headman, he told me the same: "Everybody should celebrate together." The Rana villagers in Geti did not insist on celebrating Holi according to their traditions, but in other villages there was more confusion and disagreement.

In some Rana-dominated villages south of Dhangadhi, the village headmen could not make up their mind whether to start Holi or not. A meeting was held in Hari Lal Rana's house. Although Hari Lal has no formal position, he has great influence within the Rana society. So when he said that "Holi is not until next month" and pointed out that the Ranas ought to celebrate together with the rest of the country, none of the headmen protested. There was no Holi celebration in Kailali until one month later, but in some Rana villages in Kanchanpur the celebration took place according to the Rana tradition.

Contrary to the pan-Tharu ideology which rejects the hierarchical caste structure, the Ranas' strategy to optimise social status fits in with what is commonly referred to as caste-climbing. A caste-climbing strategy is uniquely related to the sub-Indian context and is also supporting the caste ideology. In this way, the Rana strategy could be seen as undermining the strategy chosen by the pan-Tharu actors and their ideology. In both of these strategies, however, the past has become an important resource. In the next chapter, I will therefore discuss how tellings of the past have become central in order to support these various strategies.

Chapter 6 Historical Consciousness in Ethnic Identity Formation

"History is the self-consciousness of a nation" (Friedrich von Schlegel quoted in Pandey 1994:1).

The really important thing about myth, Malinowski (1948) emphasised, is its character of a "retrospective, everpresent, live actuality (ibid.:102). Because of this, myths will commonly be used ad hoc in order to "glorify a certain group, or to justify an anomalous status" (ibid.:102). Myths may also be important in legitimising political claims, as well as in the formation of collective identities.

There is a multitude of Tharu myths of origin and tellings⁽¹²³⁾ of the past. Among the Tharus in Dang, for instance, there are several tellings. One of them describes the creation of the first Dangora (cf. McDonaugh 1984a; Krauskopff 1989a). Another and more commonly known myth, is the one already mentioned in Chapter 5, which illustrates the social division between priest and client and explains the whole ritual system among the Tharus of Dang. This myth which explains the origins of the priestly clans (gotyar) also explains how the valley of Dang was divided into different ritual "countries" (des).

One of the myths commonly known among the Kochila Tharus in Saptari, traces their origin back to one of the popular Hindu gods, Lord Shiva. The myth tells how, when Parvati, Shiva's wife, was serving her husband food, a bundle of children fell out of from her saree (kocha). After Shiva blessed these children, they were named Kochilas after Parvati's dress and thus became the first Kochilas.

In Chapter 2, I mentioned two local myths in Kailali. In both of these myths, Ranas and Dangoras made claims to a twice-born Thakur status. The Dangoras traced their origin to the Thakurs in Ayodhya and to King Ram Chandra, whereas the Ranas claimed descent from the former kings and rulers of Rajasthan, the Rajputs. Another Dangora myth I was told in Geti was related to one of the particular priestly clans, the Dahit. This myth claims an origin to the Ahir (cowherding caste) of Ayodhya.

All these myths are locally tied to particular Tharu groups or to particular clans. They also share a common reference to important places and gods of the North Indian Hindu mythology.

In this chapter, I will discuss two myths of origin in detail. The first myth, which I hereafter will call the Buddha story, is a collective one forwarded by the Tharu Welfare Society and the pan-Tharu ideologues. It has become an important symbol in the pan-Tharu movement. The Buddha story represents a radical break with the other local myths. Stating that all the Tharus are descendants of Buddha, the Buddha story transgresses earlier jaat boundaries and makes all Tharus into relatives. In the Buddha story, the Tharus' origin is in Nepal, and linked to Lumbini, the birthplace of Lord Buddha. This Buddhist "ruin" is also important in the pan-Tharus' rejection of the Hindu caste structure and supports their claim to an indigenous Adivaasi status.

The second myth, which I hereafter will call the Rajasthan legend, is locally tied to the Ranas of the far west. This legend is forwarded by the Rana Reform Movement and traces the Ranas' origin to the Rajputs in Rajasthan. Ranas, according to this legend, are not Adivaasis of Nepal, but have a high caste origin in India. The Rajasthan legend shares, therefore, a common reference with the other local myths mentioned above.

Past and Present

Historical consciousness is something which is constructed and has to be related to the processes whereby it is formed and transformed. In periods of radical change, questions related to "where one comes from" often become important. I did not come to Nepal to study Tharu history or historical consciousness, but I realised that among the literate and politically active Tharus, accounts of the past had "seized center stage" (cf. Malkki 1995:53). The discourse on who the Tharus are, has recently changed from being a discourse by scholars, to become one the educated Tharu elite also engage in. Tharus, who "were formerly spoken for", are nowadays "intensely engaged in defining themselves" (cf. Friedman 1992:194). In Himal (July/Aug. 1993), there was, for instance, an article called "Faceless in History", where Tej Narayan Panjiyar - one of the leading pan-Tharu ideologues - attacked the representation of Tharus as a 'people without history'. The Tharus could not have hidden in the jungle, waiting for aeons, just to be discovered during the malaria eradication campaign in the 1950s. "They must", he argues, "have a history of their own" (ibid.:20-21).

History has become a means for asserting the identity of a community and also for identifying its rights and status. This has again created the needs to "narrativise events" (Pandey 1994:151). The various Tharu narratives are, to paraphrase Malkki (1995), not only descriptions of the past but rather reinterpretations of the past, and they are both moral and political in nature. Before I discuss these two myths and their main messages, I will introduce the central parts of the written sources and the ways Tharus have been described and represented in written documents. In their attempts to redefine their identity and status, as well as in their claims to certain rights and privileges, Tharus now go back to the old written sources. Their own interpretations and tellings of the past are, to a great extent, based upon the way they were depicted in a western discourse.⁽¹²⁴⁾

"What I write is the account I believe to be true. For the stories the Greeks tell are many and in my opinion ridiculous" (Hecateus).

In all the literature about the Tharus that I have come across, the Rajasthan legend has been mentioned as "the Tharu myth of origin". This legend tells about their mythic origin in Rajasthan - descendants of Rajput princesses who, accompanied by their servants, ran away from Rajasthan during the Mughal invasion. The princesses later married their servants and settled in the jungle area, the Tarai. The emphasis on this story, recounted and told by ethnographers, administrators and travellers has been represented as a collective Tharu origin myth. Later, the Rajasthan legend was specified to belong to the Rana Tharus of the far west, but also sometimes used to tell about the origin of other Tharu groups (e.g. the Chitwan Tharus).

The earlier colonial writings consist of reports and articles by colonial administrators and travellers (Hamilton Buchanan 1819; Nesfield 1885; Crooke 1896, 1897; Nevill 1904; Lévi 1905). Eager to record miscellaneous "tribal" customs and questions of history, race and origins (cf. McDonaugh 1984a), most of these reports commented in detail upon the "Rajasthan legend", which was based upon accounts made by the Tharus themselves as well as by their neighbours. This legend was considered as just another way of caste-climbing, and Nesfield, for instance, points out that "the fiction of having come from Rajputâna was invented by some of the clans merely to raise themselves in their own and their neighbour's estimation. There is scarcely any hunting tribe or caste in Upper India which has not set up a similar claim" (Nesfield 1885:33).⁽¹²⁵⁾ These early written sources about "Tharu customs and origin" did not always specify the particular location of the various communities labelled 'Tharu', which in many cases were considered an homogenous entity.

The first time the term Tharu is mentioned is by the Muslim geographer Al-Beruni (1033). "Farther on the country to the right is called Tilwat,⁽¹²⁶⁾ the inhabitants Tharu people of very black colour and flat-nosed like the Turks" (quoted in Ahmad 1993:96). Later Sylvain Lévi translated Tharu from the Tibetan "tharu-i-brgyud", meaning "the lineage/descent-group of the Tharus". His translation of a sixteenth century history of Buddhism links the Tharus to the former kingdom of Champarna.⁽¹²⁷⁾

Due to the climate and malaria, the Tarai became a place of mystery and danger that most people avoided. Some of this mystery, Guneratne (1994) suggests, has probably attached itself to the few human groups living there. An exotic account by Sylvain Lévi, who travelled through the continent at the end of last century, has to a great extent remained the attitude towards 'the Tharus' even today:

We are in the jungle, the terrible jungle, strange and mysterious under this bright moon, which, without diminishing the thick shadows, illuminates the contours. At midnight it is so cold that my coolies stop. On the road they light a fire, and nearby they roast themselves with delight. I take advantage of this (break) to move around a little. To the right as well as to the left, two elephants tied to some trees continue - with lots of noise - their neverending mastication of leaves. In the freezing cold night, they are like supernatural silhouettes which accompany the vibration of torn leaves. We are with the Tharus, the enigmatic Tharus. Behind the fences of halm, symbolising a wall of private life, a woman is singing a neverending song, accompanying herself by a tambourin. In this cold weather, at midnight, what is she doing? An adoration, an incantation? How can we know? Enclosed in their retreat, where nobody can disturb them, the Tharus never busy themselves with whatever happens in their neighbourhood. They live by themselves, hiding to the extent of secrecy their language, talking to strangers only in Hindustani (Sylvain Lévi 1905: 309, Vol. II. My translation).

There is no agreement among scholars as to the etymological source of the term 'Tharu', nor to the origin of the Tharus. One of the many views expressed in British Gazetteers, is that Tharu is derived from the Hindi word thar, meaning jungle. Since the Tharus are known as a people who lived in the rain forest, the translation 'man of the forest' is tempting. Nesfield, however, pointed out that the term is probably derived from the dialect of the tribe itself and not from a Sanskrit source (Nesfield 1885:3). Krauskopff (1989a) suggests that the term 'Tharu' was given by other more powerful groups living in "richer" civilisations, to the people living in an area most travellers were afraid of entering. She further points out that among the Dangora Tharus, 'Tharu' means man, human being. McDonaugh (1984a) also suggests that the word Tharu may have had this wider meaning as 'man', which later was applied to the group as a whole.⁽¹²⁸⁾

When it comes to the origin of the Tharus, Majumdar (1944), after bloodgroup tests and anthropometric studies concludes that the Rana Tharus are of a Mongoloid racial type. Early historical references place the western Tharus east of their present habitat, and they are associated with the Mongolian Koche and Meche people of Assam, North-East India (McDonaugh 1984a). Crooke (1896), for instance, argued that the Tharus "have undoubtedly some affinity with the Mongoloid tribes of Central Asia through Nepal" (ibid.:16). The Mongolian features of the Tharus is mentioned in much of the written material. Krauskopff (1989a) points out that they have a mixed physiognomy and speak different Indo-Aryan dialects dependent on regions and clans. Their racial features are more or less Mongolian, she adds and concludes that "they have lived in the Tarai area for a long time, maybe for ever" (ibid. 1989a, my translation).

In their studies of the Tharus, anthropologists like Majumdar (1944), Hu (1955, 1957) and Srivastava (1958) focused on the effects of Hinduism and cultural contact. These studies followed the dominant discourse in Indian sociology at that time, i.e. the relation between "tribes" and "castes", and were all done on Tharus on the Indian side of the border.

The famous Nepalese anthropologist Dor Bahadur Bista was the first one to describe the Tharus of Nepal in his book People of Nepal (1967). Then followed Rajaure (1981), McDonaugh (1984a) and Krauskopff (1989a), who all focused on the Tharus of Dang, the largest of the western Tharu groups. McDonaugh's thesis is a descriptive analysis of the social organisation, myth and ritual of the Dangora Tharu of Dang valley. Krauskopff focuses on the relationship between hereditary priests and their clients, on which, she argues, the Dangora social structure is built. Guneratne (1994), an American anthropologist who has recently finished his Ph.D. on Tharus from Chitwan, focuses on subjects such as ethnicity, class and the state. Hasan (1992) has written the most recent account of the Rana Tharus of Kheri district in Uttar Pradesh, India. So far, no major anthropological work has been written about the Rana Tharus of Kailali and Kanchanpur (Nepal), but several anthropologists have been involved with research in these areas lately.⁽¹²⁹⁾

When it comes to the different Tharu groups settled in the eastern part of the Tarai, they have - to quote Ardener (1989:215) - been very "remote" anthropologically, and nothing is so far written about them. The national Tharu elite consists mainly of Tharus from this part, Tharus who now challenge earlier versions of Tharu history. To them, history seems to have become "the very stuff of politics" (cf. Kapferer 1988), related to questions of origin, race and religion.

The Unifying Function of Myths

"[H]ave not all races had their first unity from a mythology that marries them to rock and hill?" (W. B. Yeats).

Historicism - the creation of historical traditions that justify present practices and beliefs - is an important feature of many contexts of ethnicity (Hobsbawm and Rangers 1984). The resource inherent in the past is also pointed out by De Vos (1975) who argues that a "sense of common origin, common beliefs and values, a common sense of survival - in brief, 'a common cause' - has been of great importance in uniting men into self-defining in-groups" (ibid. :5). Ethnicity, and attempts to explain group origin is thus often legitimated by myths of origin. "Origin myths establish who one is, and, because of one's progenitors, with which group one has rights and obligations" (De Vos and Romanucci-Ross 1975:365). Eriksen (1992) points to a central aspect when it comes to myths of origin: They may be boundary maintaining as well as boundary transgressing, boundary maintaining in the sense that they continue to mark differences between Us and Them, and boundary transgressing in the sense that they may exceed the boundaries and include Others into a We.

History and the discourse about the making of history is positional. It is dependent upon "where one is located in social reality, within society, and within global processes" (Friedman 1992:194). I already mentioned Tej Narayan Panjiyar's article (p. 151), where he challenged the "conventional" representation of Tharu history. In this article, Panjiyar, who himself is a Tharu from one of the eastern Tarai districts, launched another explanation about Tharu origin, arguing that Tharus are descendants of Buddha, who was born in Lumbini. Because of their religious faith, Tharus were later persecuted by "Brahmanic forces" and took refuge in the jungle. The same arguments are also forwarded in a book called "The Real Story of the Tharus" (1993) written by one of the most influential Tharus, Ramanand Prasad Singh (cf. Chapter 5). In this story, Singh argues that Tharus are descendants of Buddha (Buddhako santaan) and that they are the indigenous people of the Tarai. Singh gives several arguments for his Buddha theory: The inscription on the Ashoka pillar in Lumbini resembles "Tharu language", which proves that Buddha was a Tharu. There are also similarities between the Tharu language and Pali, which, according to Singh, is the ancient language of Sri Lanka.⁽¹³⁰⁾ The facial features of the Tharus resemble old stone images of Buddha, who, Singh argues, was not of Aryan origin. This indicates that Tharus are descendants of Buddha and of Mongolian origin, and their Mongolian origin also proves that the Tharus were originally non-Hindus. Even today, Tharus are only Hinduised on the surface, and there is allegedly more Buddhism than Hinduism in their religious practice, Singh argues.

Although Singh refers to various written sources in order to support his theory, the Buddha theory is his own invention and not representative of Tharu thinking. It has nevertheless been embraced by the politically more active Tharus, who now find it meaningful to represent themselves according to the message of this myth.⁽¹³¹⁾ Several of the authors referred to belong to the category "colonialist writers" already described above. Brian Hodgson (1849), according to Singh, pointed out that malaria resistent tribes "gained" their resistance after living in the area for at least 3000 years. Other scholars (Rhys Davids and Vincent Smith), emphasised that Buddha belonged to a Mongolian tribe, and Sylvain Lévi (1905) mentioned that King Ashoka's birthplace was "Tharuwat". Singh also mentions an article called "The History of Buddhism", where similarities between Tharu and Buddhist rituals are described, as well as the Indian anthropologist Majumdar (1944) who pointed to their Mongolian origin.

In Singh's "The Real Story of the Tharus", which was originally presented at a press conference in Bihar, he tries to "eradicate some of the mischievous and twisted, much maligned stories about the Tharus" (ibid.:6). What he wants to challenge, is the "Rajasthan legend", which - according to Singh - is "nonsense".⁽¹³²⁾ The Tharus originated in Nepal and are the descendants of the Sakyas and the Kolyas, two branches of the former Kingdom of Benares.

Let me introduce the readers to some parts of the story itself, as it is told by Singh.

The Tharus are of Nepalese origin, and they originated in and around Kapilvastu. Today Tharus are found all over Nepal's Tarai as well as in many of the borderlying districts of India, from UP to Bengal (ibid.:4).

In addition to the more standardised frame of origin myths, such as a royal glorious past followed by defeat, departure and settlement elsewhere (cf. Ortner 1989; Bates 1995), there seems to be another and more important message Singh wants to communicate: namely that the Tharus are ancestors of the first Buddhists and thus aboriginals of the Tarai area. Incorporating all Tharus, the Buddha story is a radical break with earlier Tharu tellings of the past - tellings which were locally tied to the particular group. By tracing Tharu history back more than 2,000 years, all Tharus become relatives - in other words people with whom one has rights and obligations.

"The discourse of history as well as of myth is simultaneously a discourse of identity; it consists of attributing a meaningful past to a structured present" (Friedman 1992:194). One important question to ask, Friedman adds, is where the attraction lies in making histories. Which are the historical contexts that make some symbols more attractive and powerful than others? The reorientation of the Tharu past must be related to the present political context of Nepal, which has made it, to paraphrase Ortner (1989), materially feasible, politically practical, as well as morally powerful to do so.

The Buddha myth is a narrative which fits into the current context of democratic Nepal. Democracy has led to a mushrooming of ethnically and religiously based organisations which define themselves in opposition to Hinduism and the Brahmanic ideology (cf. Chapter 5). As already mentioned, Ramanand Prasad Singh is connected to such an organisation, the Janjaati Mahasangh ("The Federation of Nationalities") which recognises itself as a "fundamentally Non-Hindu movement" (cf. Fischer 1993). A Hindu, according to one of the main Janjaati activists, is understood as one who accepts a place in the Hindu hierarchy. Janjaati is equated with the concepts non-Hindu and indigenous, and it is strongly felt and articulated by the Janjaati activists that the Janjaatis were in Nepal before the Aryans brought Hinduism and "intruded upon the indigenous cultures" (Gurung 1994, see also Magar and Tamang 1996). The Janjaati activists are against the division of people into social categories based on Brahmanic ideals, and define themselves as outside the whole varna system. "All minority groups are Janjaatis if they have not joined the mainstream Hindu national culture", the chairman of the Janjaati Mahasangh pointed out, and he repeated that Janjaatis are non-Hindu, non-varna and non-Aryan.

Buddhism, it seems, has become the antithesis to everything associated with caste and the hierarchical varna ideology. In this way, the Janjaati- movement is similar in nature to anti-Brahmanist reform movements, such as the Dravidian Movement in South India. In their attempt to meet the challenge from the west and be more in terms with western democratic ideals, these movements often tried to abolish the hierarchical caste system and therefore be - according to "western" values - more democratic.

The Dalit movement initiated by Ambedkar in post-Independence India, used conversion to Buddhism as a means to escape from their low status. This Dalit movement which rejected the whole caste system, started among untouchables in Madhya Pradesh, and united the various untouchable groups around the common identity Dalit. Many untouchables, who earlier had tried to improve their social status by strategies which supportod the caste system, such as "Sanskritisation" and claim to Kshatriya rank, converted to Buddhism. In the years following independence, the number of Buddhists increased considerably, and most of them were recruited from among the untouchables. The Dalits claimed better representation in the new and democratic order of independent India and formed a political party, known as the Republican Party. Although Ambedkar did not succeed in his claim for separate electorates for the Dalits, a special reservation system was made, where untouchable representatives were guaranteed one in seven seats in central and state legislatures (see Kolenda 1978).

The emphasis made by the Janjaatis on indigenousness makes this movement comparable to the various religious and social movements termed Adivaasi movements in India. The concept Adivaasi, which is the South Asian equivalent to "original inhabitant", was used to distinguish "tribal" from Hindu communities. The concept, Bates (1995) argues, owes a vital debt to colonial prejudice, and is in many ways a product of orientalism. Adivaasi communities cannot easily be distinguished from Hindu peasant communities by their way of life, and many of the so-called Adivaasis were once regarded as Hindus. Because of a decline in their economic position in society, or because, more recently, it has been to their advantage to call themselves Adivaasis, they became known as Adivaasis . The Adivaasi movements which started among the socalled "tribal" and other "backward" communities, can be seen as a response to the colonial government and to the Hindu settlers and moneylenders who came in their wake (Bates 1995:115). An assosciation called the Adivaasi Mahasabhaa, for instance, became an important political voice for the Adivaasis in the 1950s. This association raised claims rooted in terms of universal and natural rights. The Adivaasi movements have several aspects in common with other political movements of the nineteenth and twentieth century.

The Adivaasi identity, Bates stresses, has been forged in common experiences that are recent in origin. Instead of considering the Adivaasis as "original inhabitants", they may rather be seen as the "recently disposessed" (ibid.:119). The definition of original or anterior inhabitants is often a preliminary to claims for political or economic power. Therefore, Bates argues, rather than asking "who the Adivaasis were", it might be more fruitful to look for who it is that wants to define them (ibid.)

Both the Dalit and Adivaasi movements are examples of how Buddhism and indigenousness can be used politically. Both were escape mechanisms from a low caste status, but functioned also as political platforms. The Janjaati movement with its politicisation and explicit attack upon the caste system in the new democratic Nepal, can thus be said to synthesise the basic characteristics of the Indian Dalit and Adivaasi movements.

Buddhism and Hinduism are in Nepal strongly intertwined and in many places without clear cut distinctions. Many people often worship in both Buddhist and Hindu temples and make pilgrimage to the same religious sights. The very inclusive nature of Hinduism (and also Buddhism) has made it difficult to say when one religion starts and another begins (cf. Fuller 1992; Searle-Chatterjee 1994). Buddhism was and still is by many considered as "a part of Hinduism", and in the Muluki Ain, Buddhists were subsumed under the term Hindu.

The Janjaatis in Nepal now make a sharp distinction between the two religions and the meanings attributed them. Their definition of Hinduism, it is important to note, is based on the Brahmanic ideology of purity/impurity. Their understanding of Buddhism, on the other hand, is based upon a "western" definition of Buddhism as a religion based on egalitarian principles and thus in terms with democratic values.⁽¹³⁴⁾ The Janjaatis' emphasis on the egalitarian qualities of Buddhism could therefore be seen as an example of what has been termed "inverted orientalism" (cf. Faure 1993), a phenomenon which the author explains as the "natives'" use of western orientalist ideas of Buddhism as a "gentleman religion" (See also Borup 1995).

It is therefore in relation to the Janjaati ideology I understand the Buddha story and its break with the "Rajasthan legend". The Janjaatis and some of the Tharu actors consider it better to be outside the whole Hindu system than to have a low position within it (cf. Introduction). What is important for Singh, is the aboriginal status of the Tharus, and in order to be recognised as such, a connection to the high caste "noble" Rajputs must be avoided in favour of an equally "noble" past in Nepal. This has made it desirable and meaningful to pursue a "tribal" non-Hindu identity. The pan-Tharu actors, for instance, now describe the Tharu social structure as originally caste-less and egalitarian (cf. Chapter 5). They are also eager to find evidence to sustain their non-Hindu and "tribal" origin, as Panjiyar, one of the leading ideologues, pointed out in the magazine Himal: (July/Aug. 1993:21).

That perhaps they [the Tharus] were not originally Hindus is indicated from an order that was issued to enforce of the Muluki Ain (1854) among Tharus who lived between Morang and Dang Deukheri. Among other things, the order decrees that Tharus are not to eat pork or drink liquor, and that males are not to marry maternal cousin sisters.

Narratives and statements are one level of representation, practise another. The relationship between these two different levels of representation is rarely without contradictions (cf. Holy and Stuchlik 1983). This is also the case with the emic representation of a Buddhist past. "[A]ny theory about human life, if widely believed, will alter actual behaviour" (McNeill 1986:6). This is usually done "by inducing people to act as if the theory were true" (ibid.:6). What about the Tharu activists' behaviour - has it changed?

During political meetings and conferences, the Tharu elite's representation of themselves and Tharuness is clearly influenced by the Buddha story. Some of the Tharu activists now explicitly enact a Buddhist religious identity. Ramanand Prasad Singh has erected a Buddhist sanctuary (stupa )in his garden. And he has also attended Buddhist conferences abroad, together with other representatives from the organisation the Janjaati Mahasangh. There are, however, no Buddhist aspects in the religious life and ritual practices of the Tharus. The same religious syncretism between Hinduism and Buddhism which one finds in the valley of Kathmandu, for instance, is not present in the Tarai. Tharus in the east have for a long time lived in close contact with Nepalese and Indian Hindu caste communities (cf. Chapter 1), and due to this close contact, Tharus here have to a much greater extent been integrated into Nepalese mainstream society and influenced by mainstream Nepalese values. When I told a Tharu activist from the east that I found very few traces of Buddhism in their religious practices, his comment was: "We are Hindus in practice, but Buddhists by concept." What could he possibly mean by that? As a summarising symbol, Buddhism stands for egalitarian, democratic (modern) values and rejects the aspects of Hinduism which deal with hierarchy and stratification. Being "Buddhist by concept" may thus in an emic understanding of the concept imply that Tharus are a democratic, egalitarian and "tribal" people.

Are the Tharu activists trying to act as if the "Buddha theory was true" and thereby "inventing a tradition" (cf. Hobsbawm and Ranger 1984)? Friedman (1992) argues that it is the practice of mythologisation, rather than the realisation of myth in practice that has to be studied. Among Tharus in the eastern part, the Janjaati ideology has had stronger impact than in the far west, and the term Adivaasi is commonly used as a self-identifying concept. (Even the English term "indigenous" is used). Some Tharus in Sunsari, for instance, invited me to an indigenous festival (Adivaasi Melaa). This festival was for all Adivaasis in the eastern part of the Tarai, as well as for Adivaasis from India. This shows that the category Adivaasi may sometimes become central and thus transcend ethnic boundaries. Two young Tharus I met in Sunsari, for instance, represented themselves as "Tharu filmmakers" who had made a "Tharu film". According to them, only Tharu actors played in this film. When I had a closer look at the film-poster, I saw names which indicated that not only Tharus, but people belonging to other jaats, such as Tamangs and Newars,⁽¹³⁵⁾ also acted in the film. When I made a comment on this, they told me that it was true that other jaats also played in the film, but added that "they are also indigenous (Adivaasi) like us." In this context, the inclusive use of the Adivaasi concept - which made all Adivaasi into the same kind of people - represents a radical extension of "we-ness" in a Nepalese context (cf. Introduction).

When it comes to issues of self-identification, it is often easier to argue in negative terms by stating what one is not, than it is to state what one is. The Buddha story is useful because it clearly says that the Tharus originally were non-Hindus, but it does not say what they are today. What the Buddha story says, is that the Tharus are descendants of Buddha (and 'thereby' Adivaasi), not that they are Buddhists. A study of some of the Dalits who converted to Buddhism, showed that they did not necessarily become Buddhists. In fact, their religious worship was often more based upon Ambedkar than Buddhism (cf. Kolenda 1978). It is therefore useful to study the Buddha myth as a "practice of mythologisation" which has made it possible to organise collectively around a pan-Tharu, indigenous (and non-Hindu) identity. That the Tharu ideologues are not Buddhists "by practice" is thus not problematic as long as such a practice is not needed in order to legitimise their claims.

The Making and Breaking of Myth

"What seems true to one historian will seem false to another..., so one historian's truth becomes another's myth, even at the moment of utterance" (McNeill 1986:3).

The Buddha story represents a radical break with the "traditional" Tharu pasts. Even though the Buddha story is an elitarian discourse, not representative for Tharu thinking in general, it would not be right to consider it just a "freak utterance". The Buddha story, with its central messages: (i) Tharus are one people, (ii) Tharus are indigenous, (iii) Tharus are non-Hindus, Buddhists and non-Aryan, conveys meaning to, and has become an important key symbol for politically active Tharus. Something which also influences "ordinary people" (see p. 182).

When studying historisisation, it is important to include the different versions of "the past" (Cohn 1987b). As already mentioned, there are many tellings of the Tharu past, and the capacity of the collective Buddha story to break with other Tharu tellings is dependent upon its capacity to say something significant to Tharu actors (cf. Ortner 1989). I will therefore turn to the Ranas of the far west, and describe how they relate to this collective Tharu myth. Then I will discuss the myth of origin which is now communicated by the Rana elite in the far west. This myth goes against the Buddha story and the pan-Tharu movement.

Although very few Tharus in and around Geti were interested in talking about questions of origin, one person was pointed out to me as "the one who knows" about such isssues. This person was Hari Lal Rana, whom I introduced in Chapter 5.

Ramanand Singh and Keval Chaudhary say that Tharus are Adivaasis, and a Mongolian people who are descendants of Buddha. I do not believe in this Buddha story, because there is no proof. They say that the word Tharu is derived from Therabada, but that is wrong. Hindus were in this area long before Buddha was born. This jungle area was called thar, and Tharu is the name of the people who lived in the thar area. There are many different Tharu jaats, such as Rana, Dangora, Kochila, and Kathariya. We are not one and same jaat, but different. They both also claim that Tharus are a Mongolian people. But if they are descendants of Buddha, how can they be a Mongolian people? In the hills there is one jaat, called Thapa. They are Mongolian. Tharus are not Mongolian, and Buddha was not of Mongolian origin. Another reason why Hari Lal doubts the "truth" of the Buddha story, is related to his own understanding of Buddhism.

Buddha had his philosophy of not killing animals. A person was not allowed to kill any animal just in order to eat it. It was, however, legal to eat the meat of an animal had it died on its own. Because the Tharus, e.g. the Dangoras, always killed and ate their animals, they cannot be descendants of Buddha.

Keval Chaudhary told me once that the Kochila Tharus originally came from a place called Kotjak, which is today called Kustuntuniya. If they came from Kotjak, they cannot be Adivaasis. The Dangoras say that they are from Ayodhya and that they are descendants of the King Ram Chandra. Because of fights between Muslims and Hindus they fled northwards to Dang. If that is so, they are not Adivaasis. We Ranas at least are not Adivaasis. We came from India, from Rajasthan.

I told Hari Lal that I had heard about their origin in Rajasthan, and I mentioned the main parts of what I had come to know about the Rajput princesses and their slaves. To my surprise, Hari Lal shook his head, and told me that this was wrong (gallat kura). The Rana women, Hari Lal stressed, had never married their slaves. It is thus not only Ramanand Prasad Singh who rejects the well-known Rajasthan legend as "nonsense", and what follows is the result of many visits and interviews with Hari Lal, who started to tell me what he considers correct (thiik kura). After giving me an introduction to Indian history and the founding of the Mughal dynasty, he went on with the Rana Tharus' place of origin, Chittogarh:

The original place of the Ranas was Chittogarh in Rajasthan, the former kingdom of Mewar. About 400-500 years ago, there was a big fight between Muslims and Rajputs and our ancestors fled from Rajasthan. The Muslims were in majority, but as brave warriors the Rajputs fought fiercely. If the city was lost to the Muslims, the women were told to put the city on fire and to throw themselves onto the pyre in order to prevent them from being raped and enslaved by the Muslims. This practise, called johaar, was very common at that time in Rajasthan. Chittogarh was lost and the women threw themselves onto the pyre. There were therefore no women left who could have escaped with their slaves. Not all the men died in battle though, and some of them managed to escape. King Maharana Prataab Singh had two brothers.⁽¹³⁶⁾ One of them went to Kathmandu. The other one, together with 11 men, came to Nainital, to a place called Bharagaon (lit. "the village of twelve"). In Bharagaon, there was a Brahmin widow who was going to marry off her daughter to a Brahmin. The wedding should take place near a river. Before the Brahmin arrived, the Rajputs from Chittogarh reached the same place. The widow thought that the Rajputs were the Brahmin and his people, and she gave her daughter to one of the Rajput men. When the Brahmin came, there was a fight over the girl. As Rajputs are good warriors, the men from Chittogarh easily won the battle. Since then, the Rajputs settled in the area and started to marry local Brahmin women. Their descendants are today known as Ranas.

The Rana Thakurs are thus descendants of Rajputs and Brahmins. We were always wearing the white turban (phagya) and we always had a knife (talwar) in our belt (like Rajputs/warriors). Even today, this battle is imitated during our wedding ceremonies, when the bridegroom and his followers (barait) enter the bride's village.⁽¹³⁷⁾

To illustrate their Rajput origin, Hari Lal explained what Ranas do when they go to Hardwar on a pilgrimage.⁽¹³⁸⁾ "We Ranas do not write our names with the Nepalese priests/scholars (pandits) as other Nepalese and Tharus do. We Ranas go to Pipawali Haveli, to the Rajasthani pandits". We also take our purification baths together with people from Rajasthan and not with other Tharus and Nepalese people.

"In the beginning was the Word, and the Word was with God, and the Word was God" (Holy Bible, Genesis).

Hari Lal's telling differs from the conventional telling of the Rajasthan legend in several aspects. It argues that the Ranas are descendants of Rajputs (Kshatriyas) and Brahmins, the two highest varnas of the classical Hindu hierarchy. It also tells how the Rajputs (Ranas) (as good warriors) conquered the area from the local Brahmins. Just like a descent from Buddha "ennobled" the Tharus, Hari Lal's telling "ennobles" the Ranas. The conventional Rajasthan legend, on the other hand, does not give them an equally pure and noble status. It does not tell about any conquest because the princesses and their slaves just settled in the jungle.

The tellings of the Rajasthan legend include stories that conflict with one another. Even though there have always been contesting voices, some tellings attain greater degrees of dominance and popularity. The reason why the conventional telling of the Rajasthan legend has gained greater "dominance and popularity" is, according to Lal Bahadur, due to Mahananda Sabkota (a Brahmin) from Jhapa district:

Many years ago, Mahananda Sabkota was working at an office in Dhangadhi. During his time there, Babu Ram, a Newar had told him a lot of lies about the Rana women, e.g. that they came with their servants and later married them. He also told Sabkota that Rana women looked upon their husbands as lower in status, and therefore served them their food-plates (thaalis) with their feet. This was nothing but lies (jhutho kura). Sabkota, however, wrote it down as the Ranas' history, and it was published in a monthly bulletin called "Hamro Nepal" (lit. Our Nepal), in 1950-51 AD.⁽¹³⁹⁾ I was very angry with Sabkota, and I asked him why he had written lies about other jaats. Before writing this, Sabkota had never visited a Rana village. When he visited villages, however, he realised the mistake and said: "Babu Ram told me lies." He then wrote a new article in "Hamro Nepal" where he admitted his mistakes. Unfortunately, the first story became the well-known about the Ranas.

The written version of Sabkota coincides in several aspects with the earlier written sources about Tharus, as well as with the conventional Rajasthan telling of the legend. The dominance and popularity given this conventional telling, do not, however, make it into an Ur-text. Rather, it has to be seen as one of many selected tellings - rooted in a particular social and ideological context (cf. Ramanujan 1991).

One point to make is the greater legitimacy Hari Lal gives written texts. One aspect of the written, which many scholars generally agree upon, is the reifying quality of the past as soon as it becomes a "written record". Historical reality only comes to us in forms of textual representation (Pollock 1993). Once written down, Sabkota's story "became reality". Hari Lal has therefore started to write down his telling of the Rana history, something which may be looked upion as an attempt to "transform narrative legitimation into written documentary form" (Appadurai in Malkki 1995:252).⁽¹⁴⁰⁾ Although Hari Lal's telling differs from the conventional Rajasthan telling, they are similar in their relation to the Rajputs of North India. The Buddha story, on the other hand, which rejects any connection to the Rajputs, has emerged from a different social and cultural context. In order to understand this better, it is necessary to describe the role the Rajputs have come to play in the northern part of the subcontinent.

The Rajputs have for long been an important symbol for lower-status groups on the subcontinent who sought upward mobility (Srivastava 1958; Sinha 1962; Jenkins 1996). Since Rajputs are regarded as "the true representatives of the traditional Kshatriya class" (Sinha 1962:36), many tribes and castes have traced their origins to the Rajputs. This has been termed Rajputisation, Kshatriyasation and Thakurisation. In Rajasthan today, untouchables (Dalits), tribals and "other backward castes" (castes that are not untouchable but considered backwards in various ways) claim descent from, or historical association with the Rajput clans (cf. Jenkins 1996). Even Muslims claim Rajput descent, and this indicates, Jenkins argues, that Rajput has become an independent ideal, rather than a class within Hindu society (ibid.).

Buddhism, I showed, can be used politically. Likewise, the Rajasthan myth has political potentials. The revised telling of Hari Lal does not only establish a relation between the Ranas and the Rajputs, but establishes a direct link to King Maharana Pratab Singh, the heroic king who ruled in the 16th century and died on the battlefield while fighting against the Mughal Emperor Akbar.⁽¹⁴¹⁾

Why this emphasis on the Maharana in Hari Lal's telling? Srivastava (1958) described a similar claim made by some of the upper Rana subgroups (kuris) in Nainital, who claimed descent from the "glorious" Maharana Pratab Singh (ibid.:65). Srivastava related this to a reform movement which started among the Ranas in the 1930s, and which aimed to raise their status in the local caste system (see Chapter 5). In order to understand Hari Lal's "new" telling, I think it may be useful to look closer into the Rana Samaaj Sudhaar, the Rana reform movement mentioned in Chapter 5.

A few years ago, Diplal Rana - a well-educated representative of the Rana reform movement - was sent to Rajasthan. During his visit in Chittogarh, Diplal met with and talked to a Rajput named Durga Singh. In March 1994, together with two other anthropologists, I was also able to meet Durga Singh, who runs a shop in Chittogarh. With his big purple turban and moustache, Durga Singh looked very much like a Rajasthani postcard. When Mr Singh heard about the reason for our visit, he immediately asked one of his boys to bring a big collection of albums and started to show us photographs from the annual celebration of Maharana Pratab Singh, which takes place at Chittor fort. According to Singh, who himself is involved with the organisation of this annual celebration, chief ministers, politicians and former Maharajas attend this celebration. This celebration, the "Maharana cult", is a recent invention with clear political purposes (cf. Fasana 1996), and the main organiser is a caste association for Indian Ksatriyas, the Joint Indian Kshatriya Association (Akhil Bharatiya Kshatriya Mahasabhaa) using this occation to gather the nobility of Rajputs and Thakurs.

After having been guided through a lot of photographs from the celebration, we brought the Rana Tharu issue up. We asked Durga Singh to explain more about the Rajput princesses who, according to the legend, married their slaves and later were known as the Rana Tharus of the Tarai. Hearing this story, Mr Singh became very upset. How can people insinuate such lies? he asked and added that the Rajputs would bring to court people who told such lies. Rajput women had never married their slaves! According to Singh, there were many Rajput descendants in Nepal, for instance the Rana Prime Minister dynasty in Kathmandu and the royal family of Nepal, the Shahs. The Rana Tharus, like the Rana dynasty in Kathmandu, were descendants of the Sisodia clan of the Rajputs.

Since the Rajput social order is not so preoccupied with purity and pollution as orthodox Hindu norms prescribe, it is more inclusive (Jenkins 1996). And Durga Singh was ready to include the Rana Tharus into the Rajput fold, but not at the cost of their social prestige which is ruled by an ideology of honour/shame (ibid.). If the Rana Tharus were descendants of Rajput princesses and their slaves, the Rajput women - at some time - had "married down". Such a marriage "down" would imply lost social prestige and status for the Rajputs. It seems, therefore, that the Rana Tharus' attempt to be incorporated into the Rajput family has made it necessary to retell the conventional Rajasthan legend. Now that a contact is established between Rana Tharus and Rajputs, Hari Lal's telling has become important, and this may indicate that his telling is addressed to a Rajput audience. The choice of the Rajputs as a symbol of identification, indicates that the Rana elite's frame of reference is based upon the hierarchical ranking of social groups as it is laid down in the caste structure. The pan-Tharu ideology of ethnic incorporation, which in its inclusive way says that Tharus are the same kind of people, has no appeal among the Ranas. They exclude themselves from the other Tharu groups. Ranas think of themselves as a "caste".

It is not only the Ranas of Nepal who claim descent from this Rajput hero, and according to Srivastava (1958), such claims were also made by Ranas in Nainital. When the lower Rana sub-groups (kuri) claimed to be Thakurs, the upper kuris started to argue that they were former Rajputs of the Sisodiya clan (ibid.). Srivastava mentions five such "high" Rana kuris claiming descent from this particular Rajput clan. Some of their leaders tried to "revive an imagined Rajput heritage", and they started to call themselves Rana Thakur of Sisodiya clan, with the suffix Singh added to their names. (Singh means lion and is a common name among Rajputs and Thakurs). The result of this was that a new endogamous group emerged - the Rana Thakur of Sisodiya clan.

Recently, Amir Hasan (1992) described a similar process among the Rana Tharus of Kheri district. According to Hasan, the Ranas' claim of descent from Maharana Pratab Singh is "new". Although the conventional Rajasthan legend was commonly known and told, the Ranas had at no stage claimed descent from Rana Pratab (ibid.:217). Hasan relates this "new" claim to the activities and campaigns started by the Hindu nationalist party, the Bharatiya Janata Party (BJP) and the militant nationalist Hindu organisation, the National Volunteer Union (Rastriya Swayamsevak Sangh ). The Maharana, who is now projected as an anti-Muslim warrior, has become a symbol and upholder of Hindutva ("Hinduity/Hinduness"). Until BJP workers started a campaign and argued that the Ranas were descendants of the Maharana, nobody would claim such descent (Hasan 1992). Hasan describes a meeting in 1992, where it was decided to form a non-political organisation "to champion the Tharu cause".⁽¹⁴²⁾ With the help of BJP workers, money was collected to raise a statue of the Maharana in Chandan Chauki bazaar. Even a BJP politician came to this meeting (ibid.:218).⁽¹⁴³⁾ With its Kshatriya rule, Rajasthan is by many considered to be the protector of traditional Hindu values, and according to Jenkins (1996), Rajput nationalism has been equated with Hindu nationalism.

The relationship between Ranas in Nepal and India is very close. Chandan Chauki bazaar is just a short walk through the jungle from the village where two of the main actors of the Nepal Rana reform movement live. The Nepal Rana reform movement may therefore be part of a pan-Rana movement which has links to Rajput and Hindu nationalist forces. Hari Lal's telling, it seems, has acquired a totally different "life" than he and the initiators of the Rana reform movement originally intended, which - I suggest - was to optimise social status.

Ethnic Incorporation versus Caste-climbing?

Myths of origin often play an important role in establishing and maintaining boundaries, based on either ethnic or caste criteria. I have discussed two such myths of origin, which in many ways are in conflict. My approach to these myths has not been to sort out "true facts" from "distortions" (cf. Malkki 1995), but rather to examine what is taken to be the truth by different actors and why. "'[W]orlds made through narrations of the past are always historically situated and culturally constructed" (Malkki 1995:104). Although there is a lot of contradiction and ambiguity in this "world making", it is nonetheless in such contexts that "people act upon and riddle with meaning" (ibid.:104).

The Buddha story is the pan-Tharus' attempt to form a Tharu identity based on ethnicity, whereas the Rajasthan legend is the Ranas' attempt to improve their caste status. In the pan-Tharus' strategy, the Buddha story has become an important symbol in order to form a "joint estate" of common history and culture (cf. Eidheim 1992). The Buddha story gives them a different origin and makes them culturally "special". It rejects "Hindu" belonging as well as the inferior status Tharus were given within the Nepalese social structure. The Buddha myth "ennobles" them, it tells the story of an old and glorious past and a shared culture and heritage. The myth also acts in a potentially unifying manner - a unification which may make the Tharus politically stronger. "Buddhism" thus provides a practical means of creating an ethnic identity as well as expressing political unity and power and international contacts.

The Buddha story does not only claim their status as a separate ethnic group, but it also claims their status as an indigenous people. It has become crucial to the pan-Tharu actors that their situation is seen in an indigenous peoples' perspective. These actors have thereby linked themselves to the national indigenous movement (Janjaati Mahasangh), and they also identify with other indigenous peoples in India as well as in a more global perspective. The democratic ideology favours ideals of equality, and in the present democratic context, the actors themselves find it meaningful to define themselves as Buddhists (Buddhist is here understood as egalitarian/tribal/indigenous) and non-Hindus. Buddhism has also a much higher international reputation as a world religion than Hinduism, which is considered to be the essential feature of the "unegalitarian" Indian caste system.

The Ranas, on the other hand, who identify strongly with Ranas in India, have so far not found it meaningful to join the pan-Tharu movement. The Ranas' strategy of caste climbing is one that is more specific to the subcontinent and similar to processes which have been termed Rajputisation, Ksatriyasation and Thakurisation.

When it comes to questions of identity and status, ethnic incorporation and caste climbing represent totally different strategies. The first one is inclusive and claims that all Tharus are the same kind of people. It is thereby based upon egalitarian principles, and the actors also reject the hierarchical caste structure. The second strategy is exclusive in nature and supports the values that are inherent in the caste structure. Among the Ranas, the caste structure does not seem to be dissolving in favour of a form of organisation based on ethnicity.

The Ranas' attempt to climb the caste ladder is also linked to a current and modern ideology - that of Hindu nationalism. The Maharana has not only become a powerful ideological tool for Rajput nationalism but he has also become an important symbol of Hindu nationalist forces, which portray Rajasthan as the preserver of traditional Hindu values (cf. Jenkins 1996).

Because of ideological currents and relations to the world outside, the educated Tharu elite have come to handle a "larger and larger repertory of new and strange sign material" (Eidheim 1992:17). But, as we will see below, this is also reaching out to a wider part of the Tharu population. The Buddha story is related to global current ideologies, whereas the Rajput myth is oriented towards India. Common for both, is that they are attempts to adjust to a changed environment and to aspects of what is commonly labelled "modernisation", understood as the increased importance of the nationstate, bureaucracy, democracy and education (cf. Introduction).

"For those who have abandoned their traditional faith there is no hope of any return; because the perpetution of tradition requires above all that one should be unaware of it" (Abu Hamid Al-Ghazali, eleventh century Arab philosopher).⁽¹⁴⁴⁾

The Tharu elite's attempt to form a collective Tharu identity through a "historicist movement" is not only related to questions of origin and political organisation. In such movements, "ideology, political organization and cultural identity merge" (Eriksen 1992:28). The Buddha story has in many ways become a "handbook" into a Tharu culture for entrepreneurs who try to form a collective Tharu identity and a common understanding of the past. Through the rhetorical discourse of language and culture, the various pan-Tharu meetings are fora where these ideas are tested and communicated. It is thus not only a collective myth of origin which is important in the pan-Tharu activities. The idea that Tharus share a common cultural heritage has equally become important, and through cultural activities, the pan-Tharu actors hope to shape a historical consciousness and feeling of continuity with the past.

The emphasis on "Tharu culture" has led to a cultural revitalisation, and I will now look closer into how this notion of a Tharu culture is used rhetorically by the elite to justify their claim of being a people (cf. Chapter 5). "There is never a Tharu meeting without a cultural programme (samskrti karyakram)", the chairman of the Tharu Culture Society pointed out to me, and during these meetings, the pan-Tharu actors are asserting their cultural diacritica as well as selecting signs that can "stand for" a Tharu identity.

This Tharu cultural unity is not expressed without paradoxes and contradictions, however, and the need to emphasise unity in emotional speeches may indicate that this has not yet become a conventionalised "taken-for-granted" knowledge by the Tharu population as a whole. There is at present, no consensus as to what constitutes "Tharu culture".

During the aforementioned culture conference in Saptari (cf. Chapter 5), the president of the Tharu Language and Literature Council emphasised in his speech that: "Culture is the identity of the particular ethnic group. Without culture, Tharu existence is impossible." This rhetorical statement is very similar in nature to what Handler (1988) has described in the Quebecois nationalists' rhetoric about "heritage protection" and their search for "a folk society". The most basic assumptions of nationalist ideology "concern the existence of a geographically, historically, and culturally unique nation" (ibid.:154). Just like the Quebecois, who argue that "we are a nation because we have a culture" (ibid.:142), culture has become an important argument in the pan-Tharus' claim of being a people. In many ways, Tharu culture has become an "object of devotion" (cf. Kapferer 1988) to these actors and a symbol of their separate identity and claim to "peoplehood" (see also Guneratne 1994). The Tharu elite has become aware of "having a culture" (cf. Handler 1988), and in addition to a public accentuation of belonging to a people, or, to use Handler's expression, a "folk society", they have started to represent/express themselves in terms of "our culture" (hamro samskrti) and "Tharu culture" (tharu samskrti).

The need to protect and preserve Tharu culture, was strongly emphasised during the TWS convention in Bardiya. The delegates seemed to fully agree that Tharus now needed to "map out the culture in every district, to collect songs, dances and "old things" (purano cij). The suggestion to establish a Tharu art academy as well as a Tharu museum was also met with great support. In this Tharu museum, there should be a collection of Tharu objects from all over the Tarai. This reified conception of 'culture' as a thing which can be collected is originally a western conception of culture (cf. Keesing 1990). But, paradoxically, it has been adopted by Third World elites in their rhetorical, counterhegemonic discourse directed towards the west, or towards the state and the more powerful groups of their society. "If 'a culture' is thinglike, if cultural essences endure, then 'it' provides an ideal rhetorical instrument for claims to identity, phrased in opposition to modernity, Westernization, or neo-colonialism" (ibid.:48).

When the more culturally self-conscious Tharus have found out what it means to be Tharu "in a new way", a process of folklorisation of the Tharu culture is clearly visible.⁽¹⁴⁵⁾ The pan-Tharu meetings, where Tharu dances and songs are performed, and where the Tharu women put on their traditional dress, have several aspects of folklorisation. Upon his arrival at the TWS convention in Bardiya, Prime Minister Man Mohan Adhikari was welcomed by local Tharu girls dressed in their traditional costumes. They had "Tharu flowerpots" on their heads and presented flower garlands to the prime minister. The whole event was taped on a video camera, and during the opening ceremony, some traditionally dressed girls from Dang were singing a "Tharu song". This process whereby Tharu culture is represented as "traditional dress, dances, artefacts" (cf. Keesing 1990:53), fits well in with what Keesing has called "metonymic transformation", where the reified, essentialised culture is represented by its "fetishised" material forms and performances (ibid.:53).

The Tharu Welfare Society is currently publishing periodicals and newsletters in various Tharu languages. The TWS has also started to republish a periodical called "Tharu Culture" (Tharu Samskrti). The local elites have started to publish magazines with poetry, folk-songs and short stories in local languages.⁽¹⁴⁶⁾ In Bardiya, I met Ram Sagar Chaudhary, a "Tharu poet" from Sunsari. He is the author of a collection of "Tharu poems" called "With Great Costs: a Collection of Tharu Points" (Anamol Bihal; Tharu ko bita sangraha).

This collection mirrors the pan-Tharu rhetorical discourse, which I have described in Chapter 5. The poem "Our Tharu Community" (Hamar Tharu Samaaj) is one example of how the elite's discourse has become manifest in literary activities on a more local/regional level.

Our Tharu community
gathered in traditions
losing minds and intelligence
forefathers did not think
all young people became unemployed.
For two mouthful's of rice,
we became dependent on others.
Playing against progressiveness
to keep alive and cover our bodies
we followed others,
lost our continuity
and stopped out of fatigue.
Leaving all friends and colleagues
losing our dignity
becoming backward day by day
is this the development of Tharus?⁽¹⁴⁷⁾

This poem also reveals the ambivalence found in the Tharu elite's discourse, a discourse which is a mixture of old and new ideologies. The idea that Tharus are ignorant has prevailed among their Pahaari neighbours. The idea that the Pahaaris had the knowledge to become dominant in inter-ethnic matters, has in many ways become a "truism" and something which reflects the Tharus' self-image.

The poem starts with a critique of the "Tharu tradition and forefathers", considered as the main reason for the "backwardness" and problems the Tharus face today. Because the "Tharu community" was "gathered in tradition", and because their "forefathers did not think" (were illiterate), young Tharus were not encouraged to go to school. They thereby lost "minds and intelligence", which again resulted in their incapability to compete for positions in a national context: "all young people became unemployed."

The last part of the poem, however, takes up the problems related to assimilation. By "following others" (assimilation) Tharus "lost" their culture. And with this loss of culture, they also lost their dignity and became "backward day by day". Many Tharu communities (especially in the east) are strongly marked by adaptation to mainstream Nepalese values. The acculturated Tharu actors' perception of themselves and their culture as backward and inferior was influenced by mainstream notions and ideas which were dominant in Nepal (cf. Introduction and Chapter 1). Tharus, all over the Tarai, as I have pointed out several times, tried therefore to "undercommunicate" their Tharu identity and assimilate Pahaari values (cf. Chapter 3).

In his study of the Sami movement, Eidheim (1992) focused on the elite's role in bringing a new ethnic self-awareness to the Sami. The elite's emphasis on aspects of the Sami culture, history and language, gave substance to the Samis' conception of themselves and of their culture as something valuable and worth taking care of. The central element in this process of cultural revitalisation was a cognitive organisation of the Sami self-understanding of being a distinct people (ibid.:4). Eidheim has conceptualised this as a process of dichotomisation and complementarisation (ibid.:5. See also Eidheim 1971b:82). The Samis needed to maintain a contrasting identity in relation to the Norwegian dominant society (dichotomisation). At the same time they emphasised that their identity was complementary. This complementarisation would give them equal status in relation to the Norwegian majority. It was important for the Samis that they were conceptualised as different, but equal. Eidheim's concepts seem to be useful when it comes to the way Tharus currently build up their selfunderstanding through various cultural activities.

The Tharus' accumulation of "backwardness", which in their rhetoric has become a political asset (cf. Chapter 5), is not without contradictions. The educated Tharu elite do not consider themselves to be backwards, and seeing how the local Tharus in the west lived, the "we-ness" is quickly turned into a "they". "How backwards they are, the Tharus here", was a common statement among Tharus from the east vis-à-vis the Tharus in Bardiya.

When, for instance, the Tharus from Bara district saw that wooden wells were common among the local Tharus in Bardiya, they were quite surprised. In Bara, such wooden wells were long ago abandoned and every house now has a water pump. Tharus from the more developed eastern part of the Tarai therefore feel that they have a role in teaching the Tharus in the west, something which also reflects their social interaction.

When I went for a walk with some Tharus from Saptari district, we were invited into a local Tharu family in Bardiya, and they served us papaya. Before accepting the papaya, one of the men from Saptari told the locals that they ought to wash and clean the papaya better before serving it. Afterwards, when the papaya came back, clean and served on a steel plate, the man from Saptari told me: "I had to teach him to do it the right way, but now he knows."

During the convention in Bardiya, the delegates were frequently taking notes. This was, they told me, for their own research, because they attended to look and learn as well. While I had tea with some note-taking Tharus from Saptari and Morang, a few local Tharus came for the same purpose of drinking tea. One of the locals had an earring in the upper part of his ear, and a man from Saptari asked him why he had this earring. He pointed to his own ear and said "I also have one". The local Tharu explained that before he was born, his parents had lost a child. In order to avoid that the next child also died, the shaman (gurwa) told his parents to pierce the upper part of the child's ear. When he heard this explanation, the man from Saptari became very happy, because it was exactly the same reason given for his own pierced ear. "There you can see how our cultures resemble." The culture "is the same" (ewta kura ho), he concluded writing it down in his notebook.⁽¹⁴⁸⁾

It was not only during formal pan-Tharu meetings that I encountered culturally self-conscious Tharus. When I made a short visit to the eastern part of the Tarai, the idea of having an authentic (khaas) culture was strongly articulated by Tharus on a local village level. Many Tharus in Sunsari and Saptari thought it was better for me to do research on them, because their culture was pure, original (khaas), whereas the Tharu culture in the west was mixed (misieko). The Tharus of this eastern part of the Tarai have, for reasons outlined in Chapter 1, been much more assimilated into mainstream Nepalese culture than, for instance, the Tharus of the far west. Their customs and practices are therefore also influenced by and "mixed" with this national mainstream culture. Since quite a few of the Tharus in this eastern part are educated, they work in "white-collar" jobs, and therefore do not till their land. They are in this way the ones who are most removed from from the "traditional" way of life.⁽¹⁴⁹⁾ Still, they talk about their culture as the most authentic and "pure" (khaas): "Come here, and we will show you everything of our culture", they said to me. They also talked a lot about the richness of Tharu culture, which, according to them, is the richest culture of Nepal (sabbhanda dhani samskrti).

The idea that their Tharu customs and traditions (calan) differ from other neighbouring communities, may not be a new idea among Tharus, but the idea of their status as indigenous is a new one, and one which makes them, in their and in the international opinion, a unique and "authentic" people. This emphasis on authenticity encountered in the east may be related to their claim for peoplehood, because, as Keesing also has pointed out: "to claim that 'it is our culture' is to make claims of identity, authenticity, resistance and resilience" (1990:53).

"There is honor to be gained through cultural creativity, and the Santal elite have begun to experience the pleasure of cultural innovation per se" (Orans 1965:107).

Tharu cultural production is a highly political process which is used rhetorically as a vehicle for ethnopolitics (cf. Keesing 1989). The pan-Tharu movement is forwarded by the Tharu Welfare Society and the national Tharu elite - mainly based in the east.

There are today several other actors and organisations involved in this pan-Tharu movement, and some of these organisations, such as the Tharu Culture Society and BASE were founded by young Tharus in the western part of the Tarai. BASE is an organisation which works with grassroot mobilisation in the far west, and in their written material, BASE emphasises the indigenous status of the Tharus, although this was not explicitly emphasised in the various night classes. And instead of emphasising the Buddha story with its claim to a non-Hindu status, BASE planned to revive and write down the Dang legend (cf. Chapter 5).

Ordinary Tharus in the west did not seem to be, to paraphrase Jackson (1989) "self-consciously indigenist" as the members of the pan-Tharu elite. Both the Ranas and the Dangoras in Kailali and Kanchanpur engaged in activities which were more in line with Rajputisation, Thakurisation or Kshatriyasation (see Chapter 3).

During my last visit in Geti, however, I noticed a change among the villagers in how they related to their Tharu identity. I found that it was not only the well-educated Tharu elite who were self-conscious about their Tharuness. When I arrived in January 1995, it was right after BASE had arranged the annual culture programme at the hospital compound outside Geti village.

I have earlier mentioned the need felt by BASE to preserve the Tharu culture. Some of the financial support BASE receives from its various donors is today used to sponsor Tharu culture. This is done in the annual culture programme. The culture programme in 1995 took place outside Geti,⁽¹⁵⁰⁾ and among the guests there were many important people (thulo manchhe), ministers as well as foreign guests. And it was a big event, where BASE and Tharu culture became a focus of attention for both Tharus and Pahaaris.

After this culture programme, I noticed a change in the the Tharu villagers' own appropriation of their "Tharu traditions" (Tharu ko calan). Many of the Tharus, for instance, asked me to take photos while they were wearing their old traditional clothes. This had never happened before, but I guess the interest outsiders showed in their tradition, which has been and still is considered "backward" (pachi sareko) and "bad" (naramro) by the dominant society, has influenced them.⁽¹⁵¹⁾ When Tharus came to the BASE office after the culture programme, the BASE staff greeted them with the Tharu term ram ram instead of the Nepali word namaste which was more commonly used by the staff members.⁽¹⁵²⁾ Now they argued that they "ought to protect their own culture" (afno samskrti bachaunu parchha), and therefore say ram ram. The name given the culture programme was "Let Us Protect Tharu Culture".

Among the guests attending this programme, there were many important people from Nepal as well as from abroad. The guests stayed and ate in Tharu houses in the neighbouring villages, and quite a few of them stayed in Geti. Drone P. Rajaure, a Nepalese anthropologist who has done extensive fieldwork among the Tharus in Dang, was also present and made a speech about Tharu culture and the need to take care of cultural traditions.

That marginal groups try to form strong group identities around the concepts of "culture" and "tradition", is a global phenomenon related to processes of modernisation (cf. Keesing 1989; Eriksen 1992, 1996). The idea that there is a Tharu culture is maybe not a new one, but the way it is brought into discourse, has definitely given new meaning to this idea. To an increasing degree, Tharus have objectified certain symbols and practices which they now consider as their culture. Culture is no longer an unreflected upon and taken for granted aspect of their existence. Culture, on the other hand has entered the sphere of "opinion" and is something Tharus now reflect upon and talk about.

An ethnic group is not simply a cultural group. It is a cultural group which is self-consciously aware of its culture (cf. Guneratne 1994). Does their cultural self-consciousness make the Tharus into an ethnic group? According to Smith (1991), it is enough if only a few members "possess ethnic self-consciousness" and claim to be members of an ethnic group. I will discuss this in more detail in the next chapter, where I will summarise the main arguments I have made throughout this thesis.

Chapter 7 From Castes to an Ethnic Group? - Some Summarising Remarks

Throughout this thesis, I have discussed the process of ethnic identity formation among the Tharus. I have tried to explore to what extent the Tharus are changing from castes to an ethnic group, and what a transformation from a caste identity to one based on ethnicity implies. My point of departure is that caste and ethnicity are forms of social organisation based on different principles.

Ethnicity is a kind of social identification where inclusion/exclusion are founded on distinctive cultural principles. A communication of cultural differences/similarities does not necessarily rank the various ethnic groups hierarchically - they are different, but complementary. 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, there is nevertheless an agreement about the ideological principles which rank some castes higher than others. Groups organised according to caste principles share a common ideology that people can and should be ranked hierarchically. 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. Both caste and ethnicity can be studied as schema for exclusion and inclusion.

As a whole, the Tharus consist of several endogamous subgroups which were classified according to principles laid down in the Muluki Ain of 1854. This social structure - based on orthodox Hindu values of purity/impurity - classified all the citizens of Nepal into hierarchically ranked categories known as jaat. Membership in a jaat was not based on subjective identification, and the term has been commonly understood as caste. Hence, the various Tharu groups were, to a great extent organised according to caste principles and each Tharu subgroup looked upon itself as distinct from the others.

Ethnic identity cannot be created from scratch, neither is it something naturally given (cf. Smith 1986; Guneratne 1994). Ethnicity, and in this case Tharu ethnicity, is a many-faced phenomenon. It exists on various levels and it comes into existence in various contexts. The idea of being part of a greater Tharu ethnic entity is a recent phenomenon which has been brought forth by a modernised and well-educated elite. Elites are often the main actors in processes of ethnic identity formation.

There are several factors which have contributed to what we see today expressed as an ethnic assertion. These factors, I pointed out, are both external and internal ones, and I described some of them in chapters 1 and 2. One important factor - maybe the major one - is the massive immigration of Pahaaris into the Tarai belt. After the malaria eradication campaign in the 1960s, a massive resettlement programme was implemented by the government. This immigration transformed the Tarai in a short time from"a fever hell" into an "ethnic melting pot". Although there exist great variations when it comes to the historical background, the natural and economical resources as well as population pattern of the the Tarai, these radical changes have affected Tharus all over the Tarai. There is a general feeling among the Tharus that they, as a group, became marginalised vis-à-vis the politically dominant Pahaari immigrants. Because of these processes, we see today a segmentary identification, where earlier boundaries between the formally distinct Tharu groups, become less relevant.

In order to get a better understanding of the processes forwarded by the elite, it is necessary to study ethnicity on a local community level. The purpose of my community study was to explore the principles used for inclusion/exclusion on a local level. Geti is a village inhabited by different castes and ethnic groups. Next to the various Tharu groups - Rana, Dangora and Kunna - it is inhabited by Pahaaris - both twice-born and occupational groups. Most of the villagers in Geti are peasants (kisaan) who cultivate at a subsistence level. There are, however, three big landowners who belong to the politically dominant twice-born groups.

Any study of social identification must be related to the social structure. In Chapter 3, I described three central units of identification. Next to the village, these were the patriline and the independent household. Membership in a patriline is not based on shared "blood", but on shared gods. If one worships the same gods, one belongs to the same patriline. In this sense, the Tharus differ from their neighbouring people, who more commonly will conceptualise kinship in terms of shared "blood" (cf. Krauskopff 1989a).

The Tharus in Geti look upon themselves as different castes and maintain their mutual "caste" boundaries. Food and marriage, for instance, are important markers of difference, and intermarriage rarely takes place. They also talk negatively about each other's dietary habits. Witchcraft is another important boundary marker - a social attribute given the Dangoras by the Ranas. Social identification in Geti is, to a great extent, related to social status and power relations between the various groups. The hegemonic status of the twice-born Pahaaris has made it meaningful for both the Ranas and the Dangoras to emulate Pahaari values and customs. Pahaaris function in many contexts as a model for the Tharus. In order to enhance their social status, the Ranas and Dangoras engaged in a "mutually antagonistic symbolic degradation" (cf. Berreman 1975) by attributing "negative Pahaari values" to each others. This emulation is similar to strategies known as caste-climbing (cf. Bailey 1970) and "Sanskritisation" (cf. Srinivas 1996). Both are strategies which support the caste ideology.

Although caste values are dominant, they are not the only criteria for social identification. Geti is numerically a Tharu-dominated village, something which is reflected in its political and administrative set-up. The Tharu villagers would, for instance, exclude the Pahaaris from formal positions and - in certain aspects - from full membership in the village community. In such contexts, Tharus found it meaningful to express a Tharu unity, based on ethnic principles which emphasised that they "were like brothers". The reason given for excluding the Pahaaris, was that "Pahaaris are cheaters". The Tharu version of the village past, for instance, has become a collective story about the cheating Pahaaris. Pahaaris were in this context turned into "Others".

Ethnic "substance" is not the only criterion given excluxion/inclusion of the Pahaaris. Length of stay in the village was also crucial when their membership in the village community was considered. This third form of social identification is linked to the village as a social community and thus neither based on caste nor ethnic principles.

The Tharu villagers excluded the Pahaaris from certain aspects of the village life, but at the same time they wanted the Pahaaris to participate more actively. This may seem as a paradox, but I see this rather as an indication of the pragmatism inherent in the Tharus' boundary marking and identity maintenance as well as the ambivalence in their relation to the Pahaaris. At the same time as Pahaaris were blamed for their cheating habits, they were also admired for their educational standard and skills. (People in Geti also identify with and admire skills which are considered to be "modern". Thus, one may argue that modernity is a fourth form of identification present at village level.) This ambivalence is also present in the more politicised pan-Tharu movement, forwarded by an educated Tharu elite. Education may be an important factor for engendering ethnic identities and ethnic consciousness, and in Part Two I discussed some of these organisations' role in the development of a pan-Tharu movement.

There are several Tharu organisations and entrepreneurs working with the aim to educate and modernise the Tharus. In addition to their focus on education, they are also attempting to incorporate the Tharus around a collective Tharu identity.

The Tharu Welfare Society (TWS) is the oldest Tharu organisation, and until recently the only organisation which brought Tharus all over the Tarai together. The Society was originally an organisation for the Tharu elite, and its impact on the Tharu society as a whole has not been very significant. The Society is nevertheless the main actor in the political pan-Tharu movement, and it represents itself as a body speaking on behalf of the Tharu population as a whole. Pan-Tharu gatherings are important fora for communicating their main messages.

The TWS's ideology is not uncontested within the pan-Tharu world, however. Many of the young Tharus are critical to the old people who dominate the Society and have formed other and competing organisations. One of these organisations is Backwards Society Education (BASE), an organisation which operates on a grassroots level in the western Tarai. BASE started as a literacy campaign among the Tharus in Dang, but it soon expanded to all the districts in the far west where the Dangoras are settled. With its focus on bonded labour, the organisation has reached international and national recognition. Although BASE was considered a communist movement by the more conservative and monarchy-friendly members of the TWS, the international support given BASE has made this organisation into an important ally for the Society. The problem with bonded labour, for instance, is used in the TWS's claim for special treatment of the Tharus as an indigenous group (Adivaasi).

The third organisation I mentioned is the Tharu Culture Society (TCS). This organisation was also initiated by young Tharus and has local branches in several districts throughout the Tarai belt. One of the organisation's aims is to make the Tharus culturally self-conscious and to teach Tharus to appreciate their cultural identity. Another aim is to unite Tharus all over the Tarai. The TCS, therefore, works to improve the contact between the various Tharu groups. During the culture conference, for instance, several marriage alliances were established between Tharus in the east and those in the west. To accept marriage-alliances in Nepal, I pointed out, is the same as to say that one is the same kind of people. These marriages were thus important symbols stating that Tharus from Mechi to Mahakali are one people, and form a community (samaaj), not a caste within the Nepalese caste structure. In 1994, the TCS arranged the first international Tharu culture conference.

The rhetorical use of the expression "Tharu culture" has become an important vehicle for the pan-Tharu movement and Tharu cultural creativity is closely connected to political objectives. Handler (1988) has pointed out how culture and the insistence on "having a culture" may be important arguments for "peoplehood". The need to take care of Tharu culture and tradition is now emphasised by the pan-Tharu actors, and in their culture-preserving attempt, Tharu culture has become a "thing" - something which is objectified. Much attention is also directed towards the language situation.

Like elsewhere on the subcontinent, the numerical strength of the Tharus has become an important argument in their political claims. Since the Nepali citizens are counted and classified according to linguistic criteria, the issue of "Tharu language" has become important and Tharus all over the Tarai are therefore encouraged to speak their mother tongue, although the various Tharu languages may be mutually unintelligible.

When it comes to cultural and ethnic consciousness, there is a great gap between elite and grassroots. Tharus in the eastern Tarai are more nationally integrated than their counterparts in the far west, something which is reflected in their values and lifestyle. They live in close contact with Maithili- and Bhojpuri-speaking neighbours, and many Tharus in these districts are classified as Maithili. In order to maintain their Tharuness, and by fear of losing their culture, self-conscious members of a "dominated minority" may forward a revitalisation of their cultural identity (see Eriksen 1992). And they will communicate their differences to their surroundings through an "array of ethnic markers" (ibid.). Professional Tharu ideologues, for instance, are now expressing their Tharu identity in a self-conscious, reflected way.

Tharus in the west, on the other hand, are geographically more "remote" from the centre, something which has made it easier for them to retain many aspects of their cultural distinctiveness, such as language and customs. Tharus in and around Geti, for instance, do not assert their Tharuness in a similar selfconscious way as the Tharu elite - which is mainly from the east. BASE and its literacy campaign and consciousness raising among the Tharu communities in the far west, however, have brought the elite's knowledge out to a wider spectrum of the Tharu population. After BASE began to organise the annual culture programmes, there has been a change in the Tharu villagers' conception and expression of identity. When Tharus in Geti saw that "important people" and foreigners were interested in Tharu culture, their own apprehension of "their culture" also changed. The fact that they were urged to "protect their culture" has engendered a new self-understanding and a new relation to their cultural identity. From being something they had often undercommunicated, their Tharu identity is now considered - in some contexts - an asset.

Through education and the distribution of knowledge, the elite's strategy is to make all Tharus more conscious about their Tharuness. The notion of a Tharu culture, therefore, plays a central role in shaping a separate Tharu identity. Through these pan-Tharu gatherings, Tharus all over the Tarai are brought together in order to learn about their cultural distinctiveness. The speeches held during such gatherings are much like educational lectures - with a purpose to instruct and teach.

Questions of origin may often play an important role in establishing and maintaining boundaries. Myths of origin, for instance, establish relations of obligations and interests. In Chapter 6, I discussed two such myths of origin - which in many ways are opposing, contrasting tellings. In the pan-Tharu actors' attempt to form a "joint estate" (cf. Eidheim 1992) a collective myth of origin has become an important symbol. The main messages of this collective myth - the Buddha story - is that the Tharus are descendants of Buddha, and that they are Adivaasi - an aboriginal people of the Tarai. It has become crucial to the pan-Tharu actors that their situation is understood in an indigenous people's perspective. The Buddha story does not only claim their status as an indigenous people, but it also claims a Buddhist and noble and thereby non-Hindu and "tribal" origin. The Tharus were classified among the lower castes - as alcohol-drinking and enslavable. What we see today manifest in their emphasis on a non-Hindu, Buddhist origin, must be understood in relation to their low position within the Hindu social order. Buddhism has been used as an escape mechanism by many low-status groups in South Asia, such as the Dalits (untouchables in India). I therefore suggested that the Buddha story and its main messages could be seen as a synthesis of the Adivasi and Dalit movements elsewhere on the South Asian subcontinent. This indigenous and "Buddhist" status opens for new alliances, and the Tharus now identify with other indigenous and "non-Hindu" peoples (Janjaatis) in Nepal. It also links them to indigenous peoples (Adivaasis) all over the subcontinent as well as to indigenous peoples elsewhere in the world.

Neither the Buddha story nor the pan-Tharu movement seem to convey much meaning to the Tharus of the far west, and I mentioned the Ranas in particular. When it comes to questions of origin, the Ranas maintain their boundaries vis-à-vis the other Tharu groups and identify more closely with Ranas in India. Their Rana reform movement, Rana Samaaj Sudhaar, is part of a pan-Rana cross-national movement, and contrary to the pan-Tharu actors the Ranas refuse an indigenous status and instead emphasise their original twice-born status within the Hindu order. However, like the pan-Tharus, the Ranas have also recently seen the need to reorient their past, and the current Rana telling of their origin represents a break with the conventional telling discussed in Chapter 6. This, I argued, had to be related to the Ranas' attempt to join the Rajput fold and the recent contact established with the Rajputs in Rajasthan. The Rajputs are not willing to include the Ranas if they claim descent from Rajput princesses and their slaves, something which would imply lost prestige. The new telling claims a direct link with the great Rajput hero, King Maharana Pratab Singh, and states that they are the offspring of Rajput men and Brahmin women. This gives them a much higher status than the conventional telling originally gave them.

The Rajput myth is more specific for the subcontinent and fits in with what has generally been explained as caste-climbing or even "Sanskritisation". The Ranas' choice of caste-climbing instead of ethnic incorporation may indicate that Ranas relate to the principles laid down in the caste structure rather than to the ethnic principles forwarded by the pan-Tharu ideologues. The link to the Rajputs represents a possibility for increased social status. In their attempt to enhance social status, the Ranas have been linked to a current aspect of modern Indian politics. Rajasthan is today considered to be the preserver of traditional Hindu values, and the heroic Maharana has become an important symbol of Hindu nationalist forces.

Both the collective Buddha myth and the Rajput myth are attempts to adjust to a changed environment. Such an attempt has linked them to current ideologies - both in a regional, national and global context.

That Ranas, in their myth of origin, distinguish themselves from the rest of the Tharus and, in a sense, also from ordinary Nepalese people, does not necessarily mean that they do so in all contexts. As pointed out in Chapter 4, the Ranas in Geti joined the Dangoras and formed - in some contexts - a joint Tharu opposition vis-à-vis the Pahaaris. The Rana reform movement also works to improve the Ranas' national integration. The need to learn Nepali, for instance, is seen as crucial in this process, something which both the Ranas and the pan-Tharus seem to agree upon. Another case which illustrates their desire for national integration, is their celebration of Holi, mentioned in Chapter 5.

The forms and contents of Tharu ethnicity in the far west is best understood as the actors' choice for what they think will be socially effective (cf. Barth 1969a). Although the pan-Tharu actors have political motives, they are also concerned with practical results. They are therefore not as dogmatic as their rhetorics may indicate. What is crucial to these actors, is how to improve their situation and status within the Nepalese nation-state. To keep on to, or create, a strong Tharu identity is therefore not contrary to the development of a national identity. It is probably the increased degree of national integration which has promoted the formation of identities based on ethnicity and region (cf. Eriksen 1993:157). The assertion of a Tharu ethnic identity thus takes place within the framework of the Nepalese state, and is not opposed to a national ideology. At the Tharu culture conference in 1994, a belonging to the Nepalese state was emphasised on several occasions. On the stage where much of the activities took place, there was a photo of the Nepalese King and Queen. During the inauguration ceremony, a candle was lit in front of this photo. The Nepalese flag was raised together with the conference flag, and the national anthem was also played (cf. Gurung and Korvald 1994).

A regional Tarai identity is not promoted by the Tharus either, and as I pointed out in Chapter 5, one of the pan-Tharus' aims is to distinguish Tharus from the other Tarai communities. Few of the Tharus are in favour of a regional Tarai-based political party, the Nepal Goodwill Party (Nepal Sadbhavana Party). During the general election in 1994, for instance, this party got a minimal support from the Tharus (cf. Gurung and Ødegaard 1995).⁽¹⁵³⁾

It is not only the Tharus who express their identities more firmly than before. This is a common feature found among many of the various population groups in Nepal. I already mentioned the ethnic umbrella organisation, the Janjaati Mahasangh and its emphasis on indigenousness.

New forms of knowledge among the Tharus, and their exposure to global and current ideologies, have made their identity into something they can talk and negotiate about. In processes of ethnic incorporation, cultural traditions, myths of origin and language are often used in order to prove that the particular group has existed since time immemorial. The Tharu actors have found "alibies" for their substantivist conception of identity in colonial writings of the nineteenth and early twentieth century (cf. Chapter 6).

To summarise, then, some of the Tharus now define themselves as an ethnic group, and try to organise themselves according to ethnic principles, whereas other Tharus, such as the Ranas, organise themselves according to caste principles. What we see as an ethnic assertion among the various castes in today's Nepal, must be related to the country's dramatic political and social changes that have taken place since 1990, but also global processes and discourses. There are various forms and aspects of Tharu ethnicity which are expressed and manifested differently, as identity formation and maintenance, and as ethnic incorporation. While some Tharus increasingly will identify themselves as an ethnic group, others may continue to conceptualise themselves as a caste. The various forms of social identification found among the Tharus, be it one based on ethnic incorporation or one based on caste-climbing, are examples of the many strategies which exist when people try to adjust to processes of modernisation and to what is commonly known as modernity.

Appendix 1 - Glossary

Nepali is the language I used most during my study, but my informants (as well as I) would frequently use various Tharu terms (Rana and Dangora). In the text, I have not specified whether the terms are Rana, Dangora or Nepali. In the glossary, however, I have indicated whether the term is Rana (R) or Dangora (D). If nothing is indicated, the term is Nepali and transcribed according to A Practical Dictionary of Modern Nepali ⁽¹⁵⁴⁾ Some of the words are Hindi or Sanskrit, although this is not specified in the glossary.

Since there is no dictionary of any of the Tharu languages, Rana and Dangora terms are transcribed according to my informants' writing, be it in the Devanagari script or in Roman letters. Some Tharu terms are also transcribed according to my own ears.

Appendix 2 - References

Ahmad, Q.
1993 [1983]: (ed.) India by Al-Beruni. New Delhi: National Book Trust.

Allen, N.
1982: "Girls. Pre-Puberty Rites amongst the Newars of the Kathmandu Valley". In Allen and Mukherjee (eds.): Women in India and Nepal. A.N.U, Monographs on South Asia.

Anderson, B.
1991 [1983]: Imagined Communities. Reflections on the Origin and Spread of Nationalism. London: Verso.

Anderson, M. M.
1988 |1971]: The Festivals of Nepal. New Delhi: Rupa & Co.

Appadurai, A.
1986a: "Theory in Anthropology: Center and Periphery". In Comparative Studies in Society and History. Vol. 28 (2).
1986b: "Is Homo Hierarchicus?" In American Ethnologist. Vol. 13, No. 4.
1993: "Number in the Colonial Imagination". In Breckenbridge, C. A. and Van der Veer, P. (eds.): Orientalism and the Postcolonial Predicament. Philadelphia: University of Pennsylvania Press.

Ardener, E.
1989: "Remote Areas". In Ardener, E. (ed.)The Voice of Prophecy and Other Essays. Oxford: Basil Blackwell.

Bailey, F. G.
1959: "For a Sociology of India". In Dumont, L. and Pocock, D. (eds.): Contributions to Indian Sociology. No. III. Paris: Mouton & Co.
1961: "'Tribe' and 'Caste' in India". In Dumont, L. and Pocock, D. (eds.): Contributions to Indian Sociology. No. V. Paris: Mouton & Co.
1970 [1969]: Stratagems and Spoils: A Social Anthropology of Politics. New York: Schocken.

Barth, F.
1960: "The System of Social Stratification in Swat". In Leach, E. (ed.): Aspects of Caste in South India, Ceylon and North-West Pakistan. Cambridge: Cambridge University Press.
1969a: "Introduction". In Barth (ed.): Ethnic Groups and Boundaries. The Social Organisation of Culture Differences. Oslo: Universitetsforlaget A/S.
1969b: (ed.) Ethnic Groups and Boundaries. The Social Organisation of Culture Differences. Oslo: Universitetsforlaget A/S.
1994: "Nye og Evige Temaer i Studiet av Etnisitet". In Manifestasjon og Prosess. Oslo: Det Blå Bibliotek.

Bates, C.
1995: "'Lost Innocents and the Loss of Innocence': Interpreting Adivasi Movements in South Asia". In Gray et. al. (eds.): Indigenous Peoples of Asia. Ann Arbor: Association for Asian Studies.

Berreman, G.
1975: "Bazar Behavior; Social Identity and Social Interaction in Urban India". In De Vos, G. and Romanucci-Ross, L. (eds.): Ethnic Identity, Cultural Continuities and Change. Palo Alto: Mayfield Publishing Company.
1979: Caste and Other Inequities. Essays on Inequality, Meerut: Folklore Institute.
1985: "Internal Colonialism and Fourth-World Movements in the Indian Himalayas". In Brøsted et al. (eds.): Native Power. Oslo: Universitetsforlaget A/S.

Bennett, L.
1983: Dangerous Wives and Sacred Sisters. Social and Symbolic Roles of High Caste Women in Nepal. New York: Columbia University Press.

Bista, D. B.
1976: People Of Nepal. Kathmandu: Ratna Pustak Bhandar.
1991: Fatalism and Development. Nepal's Sruggle for Modernization. Calcutta: Orient Longman Ltd.

Borup, J.
1995: "Zen and the Art of Inverting Orientalism". In NIAS-nytt. Dec. 1995.

Bouglé, C.
1958 [1908]: "The Essence and Reality of the Caste System". In Dumont, L. and Pocock, D. (eds.): Contributions to Indian Sociology. No. II, Paris: Mouton & Co.

Brass, P.
1974: Language, Religion and Politics in North India. Cambridge University Press.

Hamilton Buchanan, F.
1819: An Account of the Kingdom of Nepaul. New Delhi: Manjusri.

Buggeland, A. S.
1995: Kalibonga i "Cheating country". Religiøs Billedbruk og Sosial Marginalisering blant Santalene i Nepal. Hovedoppgave i Sosialantropologi, Universitetet i Bergen 1995.
1997: "Citizenship, Identity and Minority Relations. The Case of the Santals". In Gurung and Skar (eds.): Nepal: Tharu and Tarai Neighbours, Kathmandu: Bibliotheca Himalayica (forthcoming).

Burghardt, R.
1984: "The Formation of the Concept of Nation-State in Nepal". In The Journal of Asian Studies. Vol. XLIV, No. 1.

Cederroth, S.
1995: Managing the Irrigation. Tharu Farmers and the Image of the Common Good. Copenhagen: NIAS Reports.

Central Bureau of Statistics:
1995: The Population Monograph. Kathmandu.

Chakrabarty, D.
1994: "Modernity and Ethnicity in India". In South Asia 17 (special issue) pp.143-155.

Clarke, G. E.
1995: "Blood, Territory and National Identity in the Himalayan States". In Tønnesson, S. and Antlöw, H. (eds.): Asian Forms of the Nation. Copenhagen: Nordic Institute of Asian Studies/Curzon Press Ltd.

Clifford, J.
1992 [1988]: The Predicament of Culture. Twentieth-Century Ethnography, Literature, and Art. Cambridge: Harvard University Press.

Cohen, A.
1974: (ed.) Urban Ethnicity. A.S.A. Monographs, 12.
1982: "Variables in Ethnicity". In Keyes (ed.): Ethnic Change. Seattle: University of Washington Press.

Cohn, B.
1987a: "The Census, Social Structure and Objectification in South Asia". In An Anthropologist among the Historians. New Delhi: Oxford University Press.
1987b: "The Pasts of an Indian Village". In An Anthropologist among the Historians. New Delhi: Oxford University Press.

Cox, T.
1994: Backward Society Education (BASE): The Development of a Grassroots Movement. Report submitted to the Danish Agency for International Development

Crooke, W.
1896: The Tribes and Castes of the North-Western Provinces of Agra and Oudh. Delhi: Cosmo Publications.
1897: The North-Western Provinces of India. Oxford: Oxford University Press.

Dahal, D. R.
1994: "Ethnic Cauldron, Demography and Minority Politics: The Case of Nepal". Paper presented at CNAS National Seminar on State, Leadership and Politics in Nepal. June 29-30 1994.

Despres, L. A.
1975a: "Toward a Theory of Ethnic Phenomena". In Despres (ed.): Ethnicity and Resource Competition In Plural Societies. The Hague: Mouton Publishers.
1975b: "Introduction". In Despres (ed.): Ethnicity and Resource Competition In Plural Societies. The Hague: Mouton Publishers.
1975c: "Ethnicity and Resource Competition in Guyanese Society". In Despres (ed.): Ethnicity and Resource Competition in Plural Societies. The Hague: Mouton Publishers.

Deuel, P. and Meyer, K.
1997: "Who are the Tharu? National Minority and Identity as Manifested in Housing Forms and Practices". In Gurung and Skar (eds.): Nepal: Tharu and Tarai Neighbours. Kathmandu: Bibliotheca Himalayica (forthcoming).

De Vos, G.
1975: "Ethnic Pluralism: Conflict and Acommodation". In De Vos and Romanucci-Ross (eds.): Ethnic Identity, Cultural Continuities and Change. Palo Alto: Mayfield Publishing Company.

De Vos, G. and Romanucci-Ross, L.
1975: "Ethnicity: Vessel of Meaning and Emblem of Contrast". In De Vos and Romanucci-Ross (eds.): Ethnic Identity, Cultural Continuities and Change. Palo Alto: Mayfield Publishing Company.

Dogra, B.
1990: "Tension and Conflict in the Western Tarai". In Himal, Sept./Oct. 1990.

Dumont, L.
1970: "The Village Community from Munro to Maine". In Collected Papers in Indian Sociology. Paris: Mouton Publishers.
1980 [1960]: Homo Hierarchicus. The Caste System and its Implications. Chicago: University of Chicago Press.

Dumont, L. and Pocock, D.
1957: "For a Sociology of India". In Dumont, L. and Pocock, D. (eds.): Contributions to Indian Sociology. No. I, Paris: Mouton & Co.
1958: "A. M. Hocart on Caste - Religion and Power". In Dumont, L. and Pocock, D. (eds.): Contributions to Indian Sociology. No. II, Paris: Mouton & Co.

Eidheim, H.
1971a: "Entrepreneurship in Politics". In Aspects of the Lappish Minority Situation. Oslo: Universitetsforlaget A/S.
1971b: "Assimilation, Ethnic Incorporation and the Problem of Identity Management". In Aspects of the Lappish Minority Situation. Oslo: Universitetsforlaget A/S.
1992: "Stages in the Development of Sami Selfhood". Oslo: Working Paper in Social Anthropology, No. 7.

Eliade, M.
1970 [1951]: Shamanism. Archaic Techniques of Ecstacy. London: Routledge & Keegan Paul.

Eriksen, T. H.
1992: Us and Them in Plural Societies. Scandinavian University Press.
1993: Små Steder - Store Spørsmål, Innføring i Sosialantropologi. Oslo: Universitetsforlaget A/S.
1996: Kampen om Fortiden, Et Essay om Myter, Identitet og Politikk. Oslo: Aschehaug & Co.

Evans-Pritchard, E. E.
1951 [1972]: Social Anthropology. London: Routledge & Kegan Paul.
1983: Witchcraft, Magic and Oracles among the Azande. (Edited by Eva Gillies). Oxford: Oxford University Press.

Fardon, R.
1987: "African Ethnogenesis: Limits to the Comparability of Ethnic Phenomena". In Holy, L. (ed.): Comparative Anthropology. Oxford: Blackwell.

Fasana, E.
1996: "Caste and Politics in Rajasthan". Paper presented at The 14th Conference on South Asian Studies. Copenhagen: August 21-24 1996.

Faure, B.
1993: Chan Insights and Oversights. Princeton: Princeton University Press.

Fisher, W. F.
1993: "Nationalism and the Janjati". In Himal, Mar./Apr. 1993.

Friedman, J.
1992: "Myth, History and Political Identity". In Cultural Anthropology. Vol. 7, No. 2, 1992.

Fuller, C. J.
1992: The Camphor Flame. Popular Hinduism and Society in India. New Jersey: Princeton University Press.

Geertz, C.
1973: The Interpretation of Cultures. New York: Basic Books.

Gellner, D. N.
1986: "Language, Caste, Religion and Territory. Newar Identity Ancient and Modern". In Arch. europ. sociol. XXVII (1986) pp. 102-148.
1989: "Hinduism, Tribalism and the Position of Women: The Problem of Newar Cultural Identity". Paper presented at the SOAS-conference on Language, Literature and Cultural Identity in Nepal. London: April 1989.

Ghimire, K. G.
1992: Forest or Farm? The Politics of Poverty and Land Hunger in Nepal. Delhi: Oxford University Press.

Giddens, A.
1991: Modernity and Self-Identity. Self and Society in the Late Modern Age. Cambridge: Polity Press.

Gray, A. et al. (ed.)
1995: Indigenous Peoples of Asia. Ann Arbor: Association for Asian Studies 1995.

Gray, A.
1995: "The Indigenous Movement in Asia". In Gray et al. (ed.): Indigenous peoples of Asia. Ann Arbor: Association for Asian Studies 1995.

Guneratne, A.
1994: The Tharus of Chitwan: Ethnicity, Class and the State in Nepal. Ph. D thesis presented at the Department of Anthropology, Chicago University, Illinois.

Gupta, A.
1995: "Blurred Boundaries: The Discourse of Corruption, the Culture of Politics, and the Imagined State". In American Ethnologist. Vol. 22 (2), pp. 375-402.

Gurung, G.
1994: Hidden facts in Nepalese Politics. Kathmandu: Deurali Offset, Putali Sadak.

Gurung, G. M.
1997: "Migration, Politics and Deforestation in Lowland Nepal". In Gurung and Skar (eds.): Nepal: Tharu and Tarai Neighbours. Kathmandu: Bibliotheca Himalayica (forthcoming).

Gurung, G. M. and Korvald, T.
1994: "The First International Tharu Culture Conference Saptari, February 19-22, 1994". (Unpublished report).

Gurung, G. M. & Ødegaard, S. E.
1995: "Ethnicity, Ethno-Nationalism and Democracy in Nepal". Paper presented at the workshop Democracy in Asia. Organised by NIAS, IIAS, ESEAS. Copenhagen: October 1995.

Gurung, H.
1996: In Gurung, H. and Salter J. 1996: Faces of Nepal. Kathmandu: Himal Books.

Handelman, D.
1977: "The Organization of Ethnicity". In Ethnic Groups. Vol. 1, pp. 187-200.

Handler, R.
1988: Nationalism and the Politics of Culture in Quebec. Madison: University of Wisconsin Press.

Hasan, A.
1992: Affairs of an Indian Tribe. The story of my Tharu Relatives. Lucknow: New Royal Book Company.

Hastrup, K. & Ramløw, K.
1988: (eds.) Feltarbejd. Oplevelse og Metode i Etnografien. Danmark: Akademisk Forlag.

Hobsbawm, E. and Rangers, T.
1984 [1983]: (eds.) The Invention of Tradition. Cambridge: Cambridge University Press.

Hodgson, B. H.
1849: "Kooch, Bodo and Dhimal Tribes". In Journal of the Asiatic Society of Bengal. No. 18, pp. 702-747.

Holy, L. and Stuchlik, M.
1983: Actions, Norms and Representations. Foundations for Anthropological Inquiry. Cambridge: Cambridge University Press.

Hu, C. T.
1955: "Demographic Study of Village Chandanpur". In Eastern Anthropologist. Vol. IX, 1, pp. 1-20.
1957: "Marriage by Exchange among the Tharu". In Eastern Anthropologist. Vol. X, 2, pp. 116-129.

Höfer, A.
1979: The Caste Hierarchy and the State in Nepal. A Study of the Muluki Ain of 1854. Innsbruck: Universitätsverlag Wagner.

IWGIA
1993/94: "Draft Declaration as agreed upon by the members of the United Nations Group on Indigenous Population at its Eleventh Session". In The Indigenous World. Copenhagen.

Jackson, J.
1989: "Is there a Way to talk about Making Culture without Making Enemies?" In Dialectical Anthropology 14, pp. 127-143 1989. The Netherlands: Kluwer Academic Press.

Jenkins, R.
1996: "Rajput Hindutva". Paper presented at The 14th Conference on Modern South Asian Studies. Copenhagen: Aug. 21-24 1996.

Khamboo, G.
1994: State Revolution in Nepal. Kathmandu: Mundhum Publication.

Kapferer, B.
1988: Legends of People, Myths of State. Violence, Intolerance, and Political Culture in Sri Lanka and Australia. Washington: Smithsonian Institution Press.

Keesing, R. M.
1989: "Creating the Past: Custom and Identity in the Contemporary Pacific. In The Contemporary Pacific. Vol. 1, Numbers 1 and 2, pp. 19-42.
1990: "Theories of Culture Revisited". In Canberra Anthropology 13 (2) 1990, pp. 46-60.

Keyes, C. F.
1979: (ed.) Ethnic Adaptation and Identity. The Karen of the Thai Frontier with Burma. Assimilation, Ethnic Incorporation and the Problem of Identity Management. Philadelphia: Institute for the Study of Human Issues, Inc. Pennsylvania.
1982: "The Dialectics of Ethnic Change". In Keyes (ed.): Ethnic Change. Seattle: University of Washington Press.

Kittelsen, T. C.
1996: Aspects of Marriage and Change among the Rana Tharus of Western Nepal. MA-thesis submitted for the Cand. Polit. Degree. Oslo: Department of Social Anthropology, University of Oslo.

Kolenda, P.
1978: Caste in Contemporary India: Beyond Organic Society. Illinois: Waveland Press, INC.

Kooiman, D.
1996: "The Strength of Numbers. Enumerating Communities in India's Princely States". Paper presented at The 14th Conference on Modern South Asian Studies. Copenhagen: Aug. 21-24 1996.

Krauskopff, G.
1986: "Le Betail Dans la Vie Quotidienne et Rituelle des Tharu de Dang". In Production Pastorale et Societeé. No.19, pp. 83-98.
1987a: "La Couverture en Graminées de la Maison Tharu des Vallées de Dang-Deukheri. In Etudes Himalayennes. No. 1, pp. 159-163.
1987b: "La Feminité des Poissons. Un Motif Aquatique du Myth d'Origine et des Chants de Mariage Tharu (Nepal). In Cahiers de Literature Orale.
1987c: "Des Paysans Dans la Jungle. Le Piégeage dans le Rapport des Tharu au Monde de la Fôret". In Paysages et Divinités en Himalaya pp. 27-42.
1989a: Maîtres et Possédés. Les Rites et l'Ordre Social chez les Tharu. Paris: CNRS.
1989b: "Prêtres du Terroir et Maîtres de la Fôret. La Centralisation Politique et le Système Prêtrise Tharu à Dang et Deokheri". In Purusartha 12, pp. 79-100.
1990: "Les Tharu et le Royaume Hindou de Dang (Nepal). Souveraineté Divine et Endogamie Ethnique". In L'Homme 116, Oct./Dec. XXX (4), pp. 30-54.
1992: "Receptions Divines des Offrandes a l'Incarbination des Dieux chez les Tharu (Nepal)". In Purusartha 15, pp. 131-152.
1997: Corvées (Begari) in Dang: Ethno-Historical Notes. Agrarian Conditions and Social Order - Tarai Wilderness to support an Elite. In Gurung and Skar (eds.): Nepal: Tharu and Tarai Neighbours. Kathmandu: Bibliotheca Himalayica (forthcoming).

Leach, E.
1960: "Introduction: What should We mean by Caste?" In Leach (ed.): Aspects of Caste in South India, Ceylon and North-West Pakistan. Cambridge: Cambridge University Press.

Lévi, S.
1990 [1905]: Le Nepal, Etude Historique d'un Royaume Hindou. (Vol. I-III). New Delhi: Asian Educational Services.

Magar, S. A. and Tamang, P.
1996: "Ethnicity and Nationbuilding". In Himal South Asia, May/June 1996.

Majumdar, D. N.
1944: The Fortunes of Primitive Tribes, Lucknow: Lucknow University Publishers.

Malinowski, B.
1948: Magic, Science and Religion and Other Essays. Boston, Massachusetts: Beacon Press.

Malkki, L. H.
1995: Purity and Exile. Violence, Memory and National Cosmology among Hutu Refugees in Tanzania. Chicago: University of Chicago Press.

McDonaugh, C.
1984a: The Tharu of Dang: A study of Social Organisation, Myth and Ritual in West Nepal. Ph.D. thesis presented at Oxford University.
1984b: "The Tharu House: Oppositions and Hierarchy". In Journal of the Anthropological Society of Oxford. Vol. 15 (1), pp. 1-14.
1989: "The Mythology of the Tharu: Aspects of Cultural Identity in Dang, West Nepal". In Kailash. Vol. Vol. XV, No.3-4, pp. 191-206.
1994: "Tharu Mats and Social Structure". In Journal of the Anthropological Society of Oxford. Vol. XXV, No. 1, pp. 441-48.
1997: "Aspects of Social and Cultural Change in a Tharu Village Community in Dang, West Nepal, 1980-93". In Gurung and Skar (eds.): Nepal: Tharu and Tarai Neighbours. Kathmandu: Bibliotheca Himalayica (forthcoming).

McNeill, W. H.
1986: "Myth, History and Historians". In Mythistory and Other Essays. Chicago: Chicago University Press.

Mishra, H.
1990: "A Wild World that is No More". In Himal, Sept./Oct. 1990.

Mitra, S. K.
1994: "Caste, Democracy and the Politics of Community Formation in India". In Searle-Chatterjee and Sharma (eds.): Contextualising Caste: Post-Dumontian Perspectives. Oxford: Blackwell Publ./The Sociological Review.

Moerman, M.
1965: "Ethnic Identification in a Complex Civilization: Who are the Lue?" In American Anthropologist 67, pp. 1215-30.

Müller-Böker, U.
1991: "Knowledge and Evaluation of the Environment in Traditional Societies in Nepal". In Mountain Research and Development. Vol. 11, No. 2, 1991, pp 101-114.
1993: "Tharus and Pahariyas in Citawan: Some Observations Concerning the Question of Multiethnicity in Nepal". In Toffin, G. (ed.): Nepal, Past and Present. Paris: CNRS.

Neale, W. C.
1969: "Land is to Rule". In Frykenberg (ed.): Land Control and Social Structure in Indian History. Madison: University of Wisconsin Press.

Nesfield, J. C.
1885: "Description of the Manners, Industries, Religion of the Tharus and Bogsa Tribes of Upper India. In Calcutta Review. Vol. 80, No. 159, pp. 1-46.

Okamura, J.
1981: "Situational Ethnicity" In Ethnic and Racial Studies (1981), pp. 452-463.

Onta, P.
1993: "Whatever Happened to the 'Golden Age'?" In Himal, July/Aug. 1993.

Orans, M.
1965: The Santal, a Tribe in Search of a Great Tradition. Detroit: Wayne State University Press.

Ortner, S. B.
1989: High Religion. A Cultural and Political History of Sherpa Buddhism. Princeton New Jersey: Princeton University Press.

Pandey, G.
1994: The Construction of Communalism in Colonial North India. New Delhi: Oxford University Press.

Panjiyar, T. N.
1993: "Faceless in History". In Himal, July/Aug. 1993.

Pollock, F.
1993: "Ramayana and Political Imagination in India". In The Journal of Asian Studies 52, No. 2, May 1993.

Price, P. G.
1991: "Culture as a Medium of History: Kingly Models in Indian Political Behaviour". In Weideman, D. (ed.): Nationalism, Ethnicity and Political Development in South Asia. New Delhi: Manohar Publications.

Rademacher, A. and Tamang, D.
1993: Democracy, Development & NGOs. Kathmandu: Search.

Raheja, G. G.
1988: The Poison in the Gift: Ritual, Prestation and the Dominant Caste in a North Indian Village. Chicago: University of Chicago Press.

Rajaure, D. P.
1981: "Tharus of Dang: The People and the Social Context". In Kailash. Vol.8, pp. 155-187.
1993: "The Tharus: People of Nepal's Terai". In Nepal Traveller.

Ramanujan, A. K.
1991: "Three Hundred Ramayanas: Five Examples and Three Thoughts on Translation". In Richman, P. (ed.): Many Ramayanas, The Diversity of a Narrative Tradition in South Asia. Los Angeles: University of California Press.

Rankin, K.
1997: "Kamaiya Practices in Western Nepal: Perspectives on Debt Bondage". In Gurung and Skar (eds.): Nepal: Tharu and Tarai Neighbours. Kathmandu: Bibliotheca Himalayica (forthcoming).

Roosens, E. E.
1989: Creating Ethnicity. The Process of Ethnogenesis. Newbury Park: Sage Publications.

Searle-Chatterjee, M.
1994: "Caste, Religion and Other Identities". In Searle-Chatterjee and Sharma (eds.): Contextualising Caste: Post-Dumontian Perspectives. Oxford: Blackwell Publishers/The Sociological Review.

Sever, A.
1993: Nepal under the Ranas. New Delhi: Oxford & IBH Publishing Co. PVT. LTD.

Shah, G.
1996: "Backward Castes and Politics of Neo-Hinduism. A Case of Gujarat". Paper presented at The 14th Conference on Modern South Asian Studies. Copenhagen Aug. 21-24 1996.

Sharma, P. R.
1978: "Nepal: Hindu-Tribal Interface". In Contributions to Nepalese Studies. Vol. VI, pp. 1-14, Kathmandu: Tribhuvan University.

Shrestha, N. R.
1990: Landlessness and Migration in Nepal. Boulder, Colorado: Westview Press.

Singh, C.
1990: "Where Mountain Meets Plain". In Himal, Sept./Oct. 1990.

Sinha, S.
1962: "State Formation and Rajput Myth in Tribal Central India". In Man in India. Vol. 42, 1 March 1962.

Skar, H. O.
1992: "The Innocent Aid Abroad: Moral Dilemmas and Practical Knowledge. A View from the Rainforest of Nepal". In Forum for Development Studies No.1, pp. 135-148. Oslo.

Skar, S. L.
1991: "The Relevance of Hermeneutics to Anthropological Method". Appointed Lecture Topic for Doctoral Exam at the University of Oslo.

Smith, A. D.
1986: The Ethnic Origins of Nations. Oxford: Blackwell.
1991: National Identity. London: Penguin Books Ltd.
1992: "Chosen Peoples: Why Ethnic Groups Survive". In Ethnic and Racial Studies Vol. 15, No. 3, July 1992.

Srinivas, M. N.
1962: Caste in Modern India and Other Essays. London: Asia Publishing House.
1996: Village, Caste, Gender and Method. Essays in Indian Social Anthropology. New Delhi: Oxford University Press.

Srivastava, S. K.
1958: The Tharus. A Study in Culture Dynamics. Agra: University Press.

Thuen, T.
1982: "Meaning and Transaction in Sami Ethnopolitics". Tromsø: Stensilserie A, Samfunnsfag nr. 41.

Tripathi, R
1982: Political Modernization Among the Tharus After 1967 with Special Reference to Nainital District. Ph. D thesis presented at Kumanyu University, Nainital, India 1982.

Turner, R. L.
1931: A Comparative and Etymological Dictionary of the Nepali Language. London: Routledge.

Tyler, S. A.
1977: India, An Anthropological Perspective. Illinois: Waveland Press, Inc.

Unnithan, M.
1994: "Girasias and the Politics of Difference in Rajasthan; 'Caste', Kinship and Gender in a Marginalised Society". In Searle-Chatterjee and Sharma (eds.): Contextualising Caste: Post-Dumontian Perspectives. Oxford: Blackwell Publishers/The Sociological Review.

Van der Veer, P.
1994: Religious Nationalism. Hindus and Muslims in India. California: University of California Press.

Vayda, A. P.
1994: "Actions, Variations and Change: The Emerging Anti-Essentialist View in Anthropology". In Borofsky (ed.): Assessing Cultural Anthropology. Hawaii University: McGraw Hill.

Ødegaard, S. E.
1996: "Antropologi på Bortebane". In Hovedfagsstudentenes Årbok 1996. Oslo: IMA.
1997a: "'We are Hindus in Practise, but Buddhists by Concept'. Religious Reorientation and Political Mobilisation among the Tharus of Lowland Nepal". In Hovedfagsstudentenes Årbok 1997. Oslo: IMA.
1997b: "BASE and the Role of NGOs in Local and Regional Change". In Gurung and Skar (eds.): Nepal: Tharu and Tarai Neighbours. Kathmandu: Bibliotheca Himalayica (forthcoming).
1997c: "The BASE Movement and its Role in the Process of Democratisation and Development in Nepal". In Hossain F. and Myllyllä, S. (eds.): NGOs Under Challenge - Dynamics and Drawbacks in Development. Helsinki: Ministry of Foreign Affairs, Department for International Development Co-operation (forthcoming).

Tharu Publications

Singh, R. P.
1993 [1988]: The Real Story of the Tharus. Kathmandu: Tharu Samskrti.

Tharu Welfare Society

Magazines

British Gazetteers

Gazetteer Provinces of Oudh 1877, (502-507).
Gazetteer North-Western Provinces 1884, Vol XI (266-270, 370, 371, 590, 591, 600, 619).
District Gazetteer of U.P. 1904, Vol XXXIV (95-135).
District Gazetteer of U.P., Vol XXXI (101, 102).
District Gazetteer of U.P. 1905, Vol. XLII (66, 67, 74, 197).
District Gazetter of U.P. 1905, Vol. XLIV (70, 71, 128, 129).
District Gazetteer of U.P. 1907, Vol. XXXII (140-164).

Notes

1. In 1806, the Nepalese territory stretched all the way from Sikkim in the east to the borders of Punjab in the west.

2. The term jaat is similar to the Indian term jati and denotes a whole lot of concepts, such as species, sort, kind, nation, race, ethnic group, tribe, caste (cf. Turner 1931). Harka Gurung (1996) makes a distinction between the two terms in a Nepalese context. Jaat, according to Gurung, pertains more to caste divisions, whereas jati pertains more to ethnic group. I found the term jaat commonly used to denote both concepts. But, an ethnic umbrella organisation has recently chosen the term jati in order to denote "ethnic group/nationality". This is because jaat is so strongly associated with the Brahmanic caste ideology, something which this organisation wants to distance itself from (cf. chapters 5 and 6). The members of this organisation, thus, makes the same distinction between the two concepts as is done by Gurung (1996).

3. According to Van der Veer (1994), religious identities are created through ritual discourse and practice. Religious discourse is as much a discourse about the nation as about religion, he argues, and this is what he means by the term religious nationalism. Kapferer (1988), in his discussion of the religious character of nationalism, defines nationalism as a religion created as "an object of devotion"
(Kapferer 1988:1).

4. The word varna actually means colour, and was originally used to distinguish between Aryans, "the light ones" and Dashas, "the dark ones". This classsical Hindu divison of four varnas is commonly described as functions, estates and consist of brahmans (priests), kshatriyas (kings and warriors), vaishyas (traders) and shudras (farmers, workers and servants).

5. The Vedic literature is a collection of the earliest sacred literature of India. This literature, composed in archaic Sanskrit, consists of various sacred hymns/verses (veda) to deities. It is the primary source of information on Indian civilisation during the period from 1000 to 500 BC (cf. Tyler 1977). The code of Manu (2nd century AD) is essentially a treatise on social order, based on the Brahman caste's point of view. Together with the Arthasastra and the Puranaas, the code of Manu belongs to the early scientific literature on Hinduism which gives details of the Hindu synopsis

6. Over 70 per cent of the signatories of the Muluki Ain were tagadhari, e.g., Brahmins, Chhetris and Thakuris (Guneratne 1994). Høfer points out that more than 100 of the 219 members of the Court Council, which drafted the law, were Chhetris (Höfer 1979:42).

7. The reason why the Tharus were placed among the enslavable alcohol-drinkers, was, Guneratne, suggests, because they had no military imporatnce. Those classified as non-enslavable, such as Gurung and Magar, were often recruited to the army (Guneratne 1994).

8. The God Vishnu is considered to be the preserver of all which the creator, the God Brahma created in the world. Vishnu is worshipped in various forms and incarnations by devotees known as Vaishnavites.

9. The British scholar, A. M. Hocart (1950), for instance, drew upon his field experience in Ceylon, Fiji and Egypt when he compared castes with the stratification found in these countries, He also saw similarities between castes and the social organisation found in ancient Rome and Greece. (For a critique of Hocart's work, see Dumont and Pocock [1958].)

10. There is, however, no automatic correspondence between castes and varna. The varnas relate to four different functions, whereas castes are hierarchically ranked descent groups based on descent and marriage. There are, as we will also later see, often disputes about which varna a particular jaat should be identified with. Discrepancies between self-representation and representation by others are a common feature of caste-organised communities (cf. Srinivas 1962; Quigley 1994).

11. The priest (Brahmin) who, according to the varna system, is ranked highest, have a higher ritual status than the king, who as Kshatriya is ranked second (cf. Dumont 1980. See also Hutton 1946). The subordination of power to status is, according to Dumont, one of the fundamental principles of the caste system. Dumont has been thoroughly criticised for his emphasis on the subordination of power to status (cf. Raheja 1988; Searle-Chatterjee and Sharma 1994), and the discussion about kingship and priesthood in South Asia is still going on between scholars. There is, however, a general agreement that monarchical institutions are of central importance in a caste system, and that a differentiation between noble (kingly) lineages and others is commonly made (cf. Quigley 1994).

12. According to Dumont, the "disjunction between status and power" is the only suitable characteristic for comparison. We can therefore only say that there is caste where there is a subordination of power to status. "[A]ny society lacking this characteristic", Dumont requests, even if is made up of permanent and closed status groups, should be "classified under another label" (ibid.:214).

13. Except for Muslims and Europeans, the Muluki Ain, makes no distinction between "castes" and "tribals" - all were considered to be "Hindu".

14. Clarke (1995), for instance, writes about the Muluki Ain, that it was "a fixed, prescriptive taxonomy of the peoples of the territory", and that it laid out in principled form "who should eat with whom, who should marry whom, who should carry out what occupation, and how groups were to be differently punished" (ibid.:227).

15. According to Höfer, the Muluki Ain only sanctioned a caste identity, and failed to consider subjective ethnic identity. The sanctioning of caste identity, however, has probably had a reinforcing effect upon ethnic identities.

16. I am very greatful to "Madame", who let me use her house at the hospital compound in Kailali and who, on several occasions, also let me stay with her during my visits in Dang.

17. The Bhakti/Bhagat direction, which has existed since the first century BC is some form of theistic devotional religion which emphasises worship and personal devotion to a deity instead of sacrifice. These devotional movements were based on literary sources, such as the great Hindu epics and the Vedic literature. The major devotional movements accepted Vishnu, Shiva or one of the principal mother goddesses as the supreme deity (cf. Tyler 1977:101-102).

18. This was not the fact, however. Many Tharus - especially the women - had stopped drinking alcohol. Bhagats are also highly respected people.

19. At the end of my fieldwork, the head of my household told me that I now spoke Rana fluently (pura Rana), something which - I am afraid - was a great exaggeration.

20. The continual migration of people into the Indian subcontinent has made the question of antecedence difficult. The fact that the Tarai was avoided by most other Nepalese, makes it, however, easier to claim an indigenous status to the Tharus; indigenous in the sense that they were the original inhabitants before an "internal colonisation" took place (cf. Berreman 1985). Another dimension of their indigenous status is their self-defined cultural uniqueness as well as their sense of being outside the realm of the state decision making and thus open to economical, political, social, religious and racial discrimination (cf. IWGIA 1993/94; Gray et al. 1995).

21. The caste category known today as Yadav is the result of a large number of castes from different linguistic areas that, in order to strengthen their political power, came together and formed a single caste category (see Rao 1964).

22. Jimidaar/jimidaari are terms used for the tax-collecting person as well as the system.

23. When King George V was invited by the prime minister to a hunt (shikar) in 1911, there were - next to the official Nepalese party - 10,000 attendants, and 600 elephants were handled by 200 persons ( Sever 1993). The result of this hunt, which lasted for five days, was 37 tigers, 18 rhinoceros and 4 bears.

24. Since the author James Hilton launched the term Shangri-La in the 1930s, as a metaphor for the mythical country he described in his book Lost Horizon, the term has become commonly known and used. Hilton's Shangri-La was beyond time and space and ruled by spiritual harmony. In Han Suyin's novel The Mountain is Young, similar ideas are proposed about Nepal.

25. The Tharus were for a long period considered to be a malaria-resistant people. This is a truth with modifications. If a person had survived malaria as a child, he/she was immune to a renewed infection. In a survey published by the Journal of the Medical Association (1966), 71 per cent of the children in Chitwan were found to be infected with malaria up to as recently as 1956 (Müller-Böker 1991).

26. Bigha is the measurement most commonly used in Nepal. There are, however, local/regional variations of the bigha. One bigha in Kailali district in the far west equals 1.4 hectares.

27. This number is radically different from the total population number given in Dahal (1994). According to Dahal's estimations, the total population in the Tarai includes approximately 55 per cent of the total population. One reason for this difference may be that Dahal includes the hill migrants (Pahaaris), whereas they are not included in the census referred to by Gurung. The gap between these estimations says something about the rudimentary character of such census reports in Nepal.

28. A broad social categorisation of the Tarai people divides the population into 2.9 million "caste people", 1.4 million "tribals" and 1.3 million "others". The "others" are the three religious groups (Jain, Sikh, Muslim), one language group (Bengali) and people unspecified in the census (Harka Gurung 1996).

29. Written sources from the end of last century depict a jungle area which was in the process of disappearing. "The whole line of country, then, from Bareilly to Gorakhpur, was covered with forest up to a comparatively recent period" (Nesfield 1885:35).

30. The French term ethnie is often used to denote ethnic group, or what Smith (1986) defines as "clusters of population with similar perceptions and sentiments generated by, and encoded in, specific beliefs, values and practices" (ibid.:97).

31. Since the Nepali census enumerates by language, and not by ethnicity, the census figures give a misleading idea of the actual number of people in Nepal who consider themselves to be Tharu (cf. Guneratne 1994). The under-enumeration is almost total in the Bjojpuri- and Maithili-speaking eastern Tarai districts. The 1971 census, for example, recorded only five Tharus in Saptari (ibid.).

32. Based on sources from the Tharu Welfare Society, Guneratne (1994) argues that the Kochilas form the largest group all the way from Kosi zone to Narayani zone further west. My Tharu informants in the east used the general term Kochila for all the Tharus in the eastern Tarai region, but they would specify from which district they come. Another name used for Tharus both in Morang and Sunsari is Lampuchiya (lit. long-tail), a name which derives from the particular dress of the women. In Jhapa district, there are hardly any Tharu groups, but there are other indigenous groups (Adivaasi), such as Meche, Dhimal, Satars (Buggeland 1997).

33. Scholars working among the Newars in the Kathmandu valley used the term "Shrestaisation" to denote processes of social mobility among the Newars (Korvald 1994, private conversation). I use the term "Chaudharisation" on similar processes among the Tharus.

34. Shrestha (1990) uses the term Pahaarisation to describe the process whereby hill people moved down and settled in the Tarai. This settlement was, to a great extent, a government endeavour with the purpose to nationally integrate the Tarai people. Since Pahaaris were considered to be more loyal to the monarchy, they were encouraged to settle and "nepalise" the Tarai. The term Pahaarisation has therefore a double meaning: (i) in the sense that Pahaaris moved into and "took control over the area", and (ii) in the sense that Tharus have been influenced by and assimilated into Pahaari customs and values.

35. According to Mishra (1990) the malaria eradication programme of the predecessor of today's USAID - the United States Operation Mission in Nepal - was followed by a massive legal and illegal human settlement.

36. Ojha (1987), quoted in Skar (1992), argues that more than 50 per cent of the land in the Tarai went to "non-target groups", e.g., the landed hill farmers.

37. The term kamaiya is today commonly understood as "bonded labour", but kamaiya does not necessarily imply bondedness. The variations in kamaiya practices have been described by Krauskopff (1997). The kamaiya practice in Kailali and Kanchanpur is also discussed by Rankin (1997), who argues that the indigenous kamaiya practice among the Tharus in the far west was not characterised by bondedness. When I use the term kamaiya in relation to BASE's activities, it is in the sense of bonded labour. In other contexts, I use the term about landless persons who work for a landowner with whom they have signed a yearly contract.

38. The districts west of Dang-Deukheri (the far western districts) are also known as Naya Muluk (lit. new territories), a name derived from events which took place during and after the Anglo-Nepalese war (1814-16). These areas were lost to the British East-India Company, and Nepal got them back in 1858 - hence the name Naya Muluk.

39. When I first came down to this part of the Tarai in 1991 - together with the Norwegian Church Aid and its 4-wheel-driven jeep - we still got stuck in the muddy road and thereby lost the last tow-ferry of the day.

40. The examples of Malakheti and Chaumala are not unique. Very often, when too many "outsiders" moved into their villages, Ranas left and settled elsewhere. The change in settlement pattern is, therefore, the result of migration, and not because some Ranas changed their identity and "became" Pahaaris.

41. In this referendum, people voted on whether a new party-based political system should be installed instead of the partyless panchayat system. The majority was in favour of the partyless system which lasted until the "revolution" in 1990.

42. Private conversations with Krauskopff, March 1994 and with Thakur Singh Tharu, April 1995.

43. Just in the same way as the the tutelary gods of a patriline (kuldevta) can harm the members of the patriline, the gods and goddesses in the village shrine (bhuinyar ) can harm the villagers. Once a year a special offering (pooja) is made by the big bharra to satisfy the village deities and in that way maintain the prosperity of the village.

44. A shaman, according to Eliade (1970) is a magician and medicine man, a healer and miracle-worker, who may also be priest, mystic and poet (Eliade 1970:4). Although shamanism, which, to Eliade, is a "technique of ecstacy" (ibid.:4) is a religious phenomenon existing mostly in Siberia and Central Asia, the same phenomena can also be found elsewhere, for instance, in South Asia.

45. According to Krauskopff (1989a), the sanctuary of a house is, except in some very rare cases, placed in the room of the north-east corner of the house. This is considered the "highest" room, and is also the room of the eldest son, who is responsible for the ancestor cult.

46. Burghardt (1984) uses the term desa in two different meanings: One is used in the meaning "land" or "country", and another is used to denote the religious and ritual area of a people.

47. A saree is a piece of cloth, normally seven metres long, which is wrapped around the waist several times before it is finally once wrapped around the back and shoulder. Inside the saree the women usually wear a blouse which matches the colour of the saree.

48. Rajput literally means "son of a king", derived from Sanskrit, Raajan "king" and put "son". As Kshatriyas ("warriors/kings") of the classical varna system (see Introduction), Rajputs did not practice vegetarianism, a practice most often associated with the Brahmins.

49. Here it is important to note that in both these local stories, a "caste" identity is emphasised and that both consider their group to be different from other Tharu groups, something I will discuss in detail and is one of the main points of my thesis.

50. Translated from Sanskrit to English, Mahabharata means "The Great Epic of the Bharata Dynasty" and Ramayana means "The Romance of Rama" (cf. Encyclopedia Britannica).

51. When it comes to high-caste women and their life situation, this is well described by Lynn Bennett (1983).

52. When Babu Ram Chaudhary came to Geti, he started to work as kamaiya in a Rana house. With his eight bigha of land, the head of the Rana house was considered a big farmer. He is now working as a kamaiya in Kanchanpur, whereas Babu Ram has managed to get enough land to build his own house and is engaged in various systems of share-cropping. During the Communist government in 1995 and its focus on landless people, Babu Ram even got some land to cultivate as well.

53. Based upon a bonded labour survey record by BASE, Cox (1994) points to the caste-based nature of the bonded labour system (see also Chapter 1).

54. According to Ghimire, sukumbaasi is "a person who possesses neither shelter nor any apparent means of production" (Ghimire 1992:21). In recent years, however, the term is commonly used on an illegal occupant of public land (ibid.).

55. A jamindaar today may possess much more land than the legal amount of 25 bigha (cf. Chapter 1). There are several ways of doing this, for instance by having land registered in many different names. I was told that one person had registered land in the names of his kamaiyas, his dogs and even his chickens.

56. The two main Tharu groups in the far west, Rana Tharus and Dangora Tharus, use the names Rana and Chaudhary as "surnames" or jaat-identifying names. The other name(s) are given, personal names. The Kunna Tharus also have the term Chaudhary as a "surname".

57. The irrigation system in Geti and the villages around is a canal system. Inundation canals carry water to the fields. Disputes over water is a common phenomenon in such canal systems. During my stay in Geti, I never came across such disputes. The users are themselves responsible for the maintenance of the canals.

58. To get a bank loan, where the interest is only 14-15 per cent, it is necessary to pay a lot of bribes (ghuus) and to have the right know-how. The moneylender, who is from the Indian bazaar-town Paliya, comes usually to Geti after the harvest and stays in a house he has rented in the village.

59. In Nepal, as in India, the lunar calendar is most commonly used. The year is divided into lunar months which end on full-moon or new-moon days. In Nepal, they end on new-moon days. Each lunar month is divided into 29-30 days. The period of 15 days, when the moon is growing, ends with a full moon, puurnamaasi. This period is s known as "the bright fortnight", shuklapaksha. The 15 days which end with the new moon, amavasya, is known as "the dark fortnight", krishnapaksha (cf. Fuller 1992).

60. Magh sankranti is celebrated in Nepal as the changing of the winter sun which brings warmer weather. It is one of the few holidays which are not taken from the old religious lunar calendar (Anderson 1988).

61. The magni is usually established when the children are 3-5 years old. Frequently, there is a negotiator called majpatiya who also makes sure that the two parts belong to different patrilines (kurmas).

62. Dashain is an important Hindu festival which takes its topic from the Ramayana epic. Dashain, which is also known as Durga Pooja, celebrates Rama's victory over the demon god Ravana, hence it is good conquering evil.

63. Diwali is in Nepal also known as Tihaar. It is the festival of lights, and the main deity worshipped is the goddess of wealth Laxmi. Diwali usually takes place in mid-October. Laxmi pooja takes place on new moon day, but during the bright fortnight of the month kartik.

64. A ward is the smallest administrative unit of a village. A village in Nepal may be divided into several wards depending on its size.

65. The number of Pahaari houses changed from the beginning of my first stay, when I did this house counting. New houses were built, into which relatives of those already settled moved.

66. Bhui means earth, and the earth is commonly worshipped as a deity. Bhuinyar is used to denote the shrine itself, but as far as my informants told me, one of the goddesses resident in the shrine was also called bhuinyar. According to Krauskopff (1989a), bhuinyar is a term commonly used by the Dangoras in Dang, next to the Dangora word thanwa which means "the place". The thanwa in a Dangora village is placed north of the village.

67. Tharus in Geti today seem to emphasise education when they consult people. Basu Dev Chaudhary, for example, was considered the most able Tharu in the village. Since he has gone to school, he dares to speak in front of important people (thulo manchhe), the villagers emphasised, and actually Basu Dev had 11 different public positions. When I asked people in Geti to mention those who they considered powerful in the village, nobody mentioned the headman. Most matters of importance in the village today go through the ward chairmen.

68. Munro, for instance, wrote: "While the village remains entire, they [the villagers] care not to what power it is transferred: wherever it goes the internal management remains unaltered" (Munro, quoted in Dumont 1970).

69. Critiques have also argued that the village is a construct and a reification which only exists in our minds. And this reification, it is argued, prevents us from understanding the historical and contemporary situations villages have to negotiate (cf. The fourteenth Conference on Modern South-Asian Studies, Copenhagen 1996).

70. According to Basham (1954), gotra is a Sanskrit word which means cattleshed.

71. Tyler, who describes the nature and character of the Hindi gotra, comes up with the same conclusion: "The members of a clan are dispersed over many villages and one usually does not know all of one's clan fellows..., but in most cases the clan merely serves to define the limits of exogamy" (Tyler 1977:139).

72. As soon as a Rana bride has given ghee (clarified butter) to the kurma members and wedding guests, and after she has gone through the ceremony of greeting the gods in her husband's house, she is part of her husband's kurma and entitled to worship the lineage gods.

73. By extended family, I mean senior parents who live with their sons and their wives and children.

74. The kinship terms among the two main Tharu groups in the far west are very similar. Srivastava (1958) pointed out that the Rana kinship system is neither fully classificatory nor fully descriptive. The same kinship term is used for several relatives of the same generation and gender.

75. The strong position of the Rana women has been described and emphasised by most of those who have been in contact with them. I thought that I could see differences between the Rana and the Dangora women's positions. Ranas seemed to have a stronger position in the family and vis-à-vis men. But I may as well have been predisposed by all the literature I had read. and which pointed to the Rana women's strong position.

76. A Brahmin friend of mine in Kathmandu told me that whenever someone had done something stupid, it was common to say "just like a Tharu".

77. The Tharu women, on the other hand, did not envy the life of their twice-born sisters. "They have no freedom", they often said and also emphasised that the Pahaari women worked much more than themselves. The Tharu women also thought that the Pahaari women were gossiping too much. Pahaari women were not straight (siida), they said, and gave examples of how Pahaaris would smile in front of you, but laugh and gossip (kss kss garne) behind your back.

78. Inherent in the rules concerning food, the restrictions are greatest for ordinary or everyday food, known as kaccaa, such as boiled rice or roti cooked without fat. Kaccaa food (imperfect) is vulnerable to impurity and therefore reserved for own jaat. Pakkaa food (perfect food) consists of puris pancakes cooked in ghee (clarified butter) which, as a dairy product from the cow, is considered pure.

79. The system known as dowry, where the bride's part gives gifts and money to the groom, is linked to hypergamic marriage (the girl marries upwards). It has therefore been common to explain dowry as the price one has to pay to get a higher status for one's daughter. Bride-price ("to sell one's daughter") has been considered a "tribal" characteristic, opposed to the North Indian Brahmanic model of dowry and groom-price. The two main Tharu groups in the far west most commonly practice a form of bride-price, although the gifts/money flow both ways, both before and after the wedding. McDonaugh (1984a) has described these expected mutual "prestations" among the Dangoras in detail.

80. There are lots of social and structural contingencies which actually limit a Tharu woman in her freedom to leave her husband. Since a woman is usually not welcomed back to her natal home, it is necessary for her to find a new husband. She may often end up with a husband who has been left by his wife, or a man who is considered a "bad person" (naramro manche). The eldest daughter in Gorya ghar, for instance, had been married so many times that she would only end up with "bad men". This can easily become a vicious circle where a woman has to accept less and less desirable partners. Another hindrance was children. Not many men were happy to get extra mouths to feed. The patrilineal descent system makes the children part of their father's patriline. It is, however, common for a woman to take her small children with her. They may move back to their father's house when they grow older, or, in some cases, they are adopted by their mother's new husband.

81. This is usually referred to as devta dekhen, which literally means "to look at the deities". The healer has his particular deities, whom he can consult and who will give him the information needed. Usually, when the healer tries to get in contact with his deities, he will use a steel/brass plate (thaali) on which he has put some rice grains. He will move the rice grains around on the plate while chanting his mantras for some time entering into some kind of trance. This goes on for a while until the healer has "seen" the deities. He will then know what the deties asked for in order to make a person well, or the healer will offer the deities something. After the recovery of a sick person the healer will come back to the family and tell the people what they have to give the gods in a ritual offering (pooja). This may be chicken, liquor, eggs etc. The family then gives the items to the healer who in turn will offer this to his deities.

82. Later, the same boy told me how his maternal grandmother from India had died because of stomach pains caused by a witch. When I asked if he knew who the witch was, he said that it was one of the villagers (gaon ko manche). I also asked him if the witch had been a Dangora, but he said that the witch had been a Rana from the village, and he emphasised that it had been a "big witch" (thulo boksii). This indicates that there is no 100 per cent consistency in the Ranas' statements about Dangora witchcraft. In this case, what seemed important to the young boy, was to emphasise the "terrible death" his grandmother had suffered, and doing this, he "forgot" that he earlier had claimed that only Dangoras were witches. Since there are very few Dangoras in India, it is probably not so important to make a distinction between Ranas and Dangoras there. The few Dangoras in India often work for Ranas in Rana-dominated villages.

83. In the South Asian Hindu context, vegetarianism has acquired status as the superior form of diet. Tharus who have become devotees (bhagat/bhakti) of some sort are usually shown great respect. The bhagats in Geti were also among the better-off Tharu families, and the bhagat status gives them added prestige.

84. Dumont traces the history of this general hierarchy of foodstuffs back to the start of the veneration of the cow and the untouchable status given beef eaters. In intra-jaat festivitas, for instance, "perfect" food (pakkaa) is served. Weddings among Tharus are in general Tharu weddings, but if some Tharus know that they will have Pahaari guests, they will call for priests (pandits) to make pulau (rice mixed with vegetables cooked in ghee) which is "perfect" food.

85. Dangoras would not eat buffalo meat, however, because, as Hindus, this is forbidden. Rat is, on the other hand, considered a delicacy, especially among Dangoras, who go rat-hunting on the fields. The meat is considered good and clean (sapha) because the rats have been living on, and eating from, the crops.

86. In Uttar Pradesh, it is very rare to share a water-pipe or a smoke with people who do not belong to one's own jaat (Dumont 1980). Srivastava (1958) described hookah restrictions among the Rana Tharus in Nainital. Sharing of water-pipe with the lower sections (kuris) was totally forbidden, and the penalty for doing so was excommunication. These rules were drawn by a social Reform Movement in the 1930s (ibid.:106-107).

87. One of the contrasts drawn between "caste" and "tribe" was related to their relation to marriages outside the group. "Tribes", it was generally argued, would tolerate marriages outside the group, although it generally would take place within.

88. One daughter of a Brahmin landowner who married a poor boy of Chhetri caste, was thrown out of the house, and the family members consider her a dead person.

89. Amit is a very self-conscious young boy, and he is very interested in Tharu culture and traditions and has also been on Radio Nepal singing Tharu songs.

90. Strong local identities, however, do not exclude other identies. Although I did not observe many contexts where the Geti villagers expressed a national belonging, boundaries between the national and the local are blurred (cf. Gupta 1995). The general election in November 1994 was an "event" which gave me the opportunity to see that Tharus were neither encapsulated in their own reality, nor isolated from the larger unit, Nepal (cf. Ødegaard 1996).

91. The same idea that Pahaaris are cheaters exists among the Sattars (cf. Buggeland 1995).

92. Hindu is a term based on religious belonging and has nothing to do with race. It is perfectly possible to be both Hindu and Mongol, and in fact, many Nepalese of a socalled Mongoloid racial origin would recognise themselves as Hindu (cf. Gurung and Ødegaard 1995). The leaders, however, emphasise a racial and religious aspect, and argue that there is racial discrimination in Nepal, similar to the apartheid system in South Africa. The MNO has gained a lot of grassroot support among Gurungs and Limbus in the far eastern hill districts (cf. Hangen 1995, private conversation), and the leader of the MNO has presented his ideas in a book called Hidden Facts in Nepalese Politics.

93. One of the initiators to the UNF has written a book, State Revolution in Nepal, where he divides Nepal into four states - named after and based upon the indigenous population of the area (cf. Khamboo 1994).

94. When I arrived in Kathmandu in August 1993, it did not take long to grasp the politicised discourse about national, ethnic and religious issues. In the July/August volume of the well-known magazine Himal, the national project of Nepalese historiography was criticised - a project, it was argued, which had not left any space for the many marginal groups: "Nepali and non-Nepali modern-day historians of Nepal have a lot to answer for", P. Onta, one of the authors argued (ibid.:29).

95. A main critique among the young Tharu generation is that the TWS is undemocratic. They claim that the members of the central committee are appointed internally and not elected. During the TWS general assembly in Bardiya, the young Tharus wanted to suggest candidates to vote for, but the members were elected by acclamation. When he was appointed a member to the central committee, Dinesh Kumar Chaudhary declined because there had been no voting. Women, the young argued, should also be better represented. At present, there are no women in the TWS central committee.

96. The lack of an educational system during the Rana period forced many rich landowners, especially in the eastern part of the Tarai, to send their children to school in India.

97. Gopal Gurung, the founder and leader of the Mongol National Organisation, (the "pan-Mongol" organisation mentioned in the Introduction to this part), has criticised the leaders of the Janjaati Mahasangh for their choice of the term Janjaati (cf. Gurung and Ødegaard 1995).

98. It seems as though Singh treats the term jaat in the same sense as the English term "caste", which again was commonly distinguished from "tribe" (cf. Introduction).

99. The Marwaris are a caste originally from Rajasthan. They are often businessmen and, according to Guneratne (1994), well-known for their "frugal" life.

100. The Tharus of Nainital as well as those of Champaran have their own organisations. The one in Nainital, founded in 1976, is called the Rana Tharu Council (Rana Tharu Parishad), whereas the one in Champaran is called the Indian Tharu Welfare Association (Bharatiya Tharu Kalyan Maha Sangh ) and was founded in 1971 (Guneratne 1994:325).

101. One of the objectives mentioned in the TWS constitution, is to "provide consultancy services to national and international agencies and individuals" (cf. TWS pamphlet).

102. During one of my meetings with Satya Narayan Chaudhary in Kathmandu, he told me that in the eastern part of the Tarai it was the Indians who had exploited the Tharus. The Indians played the same role in the east as the Pahaaris in the west, he argued.

103. The facilities referred to are the King's appointment of Ramanand Prasad Singh as Attorney General, as well as the appointment of another Tharu, Lok Madhav Lal Tharu, as Secretary of the Home Ministry. Two other Tharus, Keval Chaudhary Tharu and Shiva Nanda Mahaton Tharu, were appointed to the Raaj Parishad (King's Council). In addition, certain seats were reserved for Tharu students in the Engineering, Agriculture and Forestry colleges.

104. In an unpublished paper, Community Associations in Indian Politics, M. Weiner discusses the role of social statistics. The fact that populations often diminish drastically from one census to another, is, by Weiner termed "genocide by census redefinition" (Weiner quoted in Geertz 1973:275).

105. This idea that Maithili has its roots in Tharu is originally one of Singh's ideas. Ramanand Prasad Singh is from a district which is within the earlier boundaries of ancient Mithila. One of his goals is to establish the Tharu language as distinct from Maithili (see Guneratne 1994:343).

106. Non-formal education (NFE) takes place outside the government-run national school system. Non-formal education therefore tries to reach those who for various reasons fall outside the government-run system and its focus on children in "school-going age". The curricula used as well as the skills taught are also in many cases different. Often non-formal education has a stronger emphasis on practical education which is suited to the particular user group in question, i.e. women.

107. According to Cox (1994), many illiterate Tharus lost their land by signing sale documents which were described to them as something else (e.g. a labour contract, or receipts for sale of grain).

108. The ruling elite, in order to secure its power, limited education to the twice-born part of the population only. The ones who belonged to the different alcohol-drinking groups (matwali) were thus excluded from education. This is still reflected in the Nepalese population.

109. This committee is working for indigenous rights in Asia, and Dilli has since 1988 been in contact with and aware of the international indigenous movement. This increased politicisation of the organisation was one of the factors leading to the break between Non-Frill's and Dilli's organisation in December 1989 (cf. Cox 1994).

110. The voluntary sector in Nepal used to be controlled by the Social Services National Co-ordination Council (SSNCC), a council created by the King in 1977 and overseen by the Queen. To be entitled to donor funding, registration with the SSNCC was necessary, and this registration had to be applied for annually. A good relationship with the powerholders was essential and any organisation with ethnic, religious or political interests was considered illegal (see e.g. Rademacher and Tamang 1993). After 1990, the SSNCC was replaced by the Social Welfare Council (SWC) - which is not controlled by the Royal family.

111. Some of BASE' s income-generating training activities are, for instance, hair cutting, weaving, carpentry, plumbing, mat weaving.

112. It is often claimed that one reason why girls are not sent to school, is because they will not remain in the family when they grow up. Girls marry out. Education is nowadays considered an important part of a girl's "dowry", which can help families in marrying daughters "upwards" (see Chapter 4).

113. The women groups consist usually of 16-17 women from the same village. The money saved from vegetable sale is put into a bank account in Tulsipur and the bank account is registered in two of the women's names. The women's group has motivation meetings with the leader of the women's programme once a month.

114. The ALC classes use a series of books, Naya Goreto, a course made for adult primers, while the children's night classes use the Naulo Bihani books written by David Walker.

115. I thought this demand would make it difficult to recruit teachers locally, but according to "Madame" there are several Tharus with eight years of schooling. Unlike many other Nepalese, these educated Tharus remain in their villages afterwards. This demand may be an ideal, and in Geti, one of the teachers had not completed class eight.

116. For a discussion on Dilli's role as a community leader, see Ødegaard 1997c.

117. All the foreign interests towards Tharu and Tharu culture made it, however, possible to have the barka naac celebration performed in 1995. The performance took place in Gorahi, Dang and was sponsored by the American couple and "Tharu students" Deuel and Meyer.

118. This legend tells about four brothers who went to Bengal to study magical formulas. On their way back to Dang, they decided to divide the valley into ritual countries (des) over which each of them would rule as a desbandhya-gurwa. After the three eldest brothers had chosen their parts of the valley, there was nothing left for the youngest brother. A quarrel started, and the three eldest brothers promised that the youngest brother would be allowed to install himself all over the valley. Although he would have the status as a priest, his priestly status was not that of the desbandhya-gurwa - linked to a country (des). Therefore, the youngest brother would have to find clients (barin) wherever he could. (For more details about the legend, see Krauskopff 1989a; McDonaugh 1984a).

119. Dinesh's father has two wives, a practice known as sauta, and which has been common among wealthy people, mainly Chhetris. The younger wife is usually called kanchi-ama, which means "the youngest mother".

120. According to Skar, the Ranas in the villages of Urma and Urmi argue that BASE is a Dangora organisation, wheras the Lutheran World Service Community project, on the other hand, is "theirs" (in a private conversation 1995).

121. After the Indian partition in 1949, the government needed land in order to resettle some of the refugees from Pakistan. Some land was therefore expropriated from the Tharus in Nainital. Other Tharus lost their land after the so-called "green revolution" in the 1970s made the area attractive for land-hungry people. (In Himal, Sept. /Oct. 1990).

122. In Nepal and India, the use of the lunar calendar system is most common (cf. Chapter 3). A lunar month, however, is shorter than a solar month, and in order to make the lunar months appear at approximately the same time every year, it is necessary to add an extra "13th" month. This is done in periods of a little less than three years and happened in 1993/94, with a postponement of the lunar months.

123. When Ramanujan (1991) discusses the many Ramayanas, he uses the word tellings instead of the usual terms "version" and "variant". This he does because the latter terms can imply that there is an invariant, an original or Ur-text (usually the Valmiki telling in Sanskrit 300 BC), which is the earliest and probably the most prestigious of them all. In order not to imply that there is an original "version" of the Tharu origin, I will therefore use the term telling.

124. The way some of the Tharu ideologues define what it means to be a Tharu and what they consider "pure Tharuness", mirrors, to a great extent, the way they were described in much of the written sources. A "pure Tharu" (suddha Tharu) , in these ideologues' opinion, is a person who lives in a village, in remote areas. This self-representation as a people who live a traditional life is similar to the western orientalist discourse and its tendency to romanticise village life. Tharu ideologues have thus taken over a language developed in the western/orientalist knowledge (cf. Keesing 1990). This paradox could be seen as a process of "inverted orientalism" as described by Faure (1993), see p. 165.

125. Nesfield (1885) adds that "Some Tharus know nothing about this tradition, and those who do are not able to tell you whether it was the sack by Alaudin (AD.1303), or that by Bahadur Shah (AD. 1533), or that by Akbar (AD. 1567)" (ibid.:33).

127. Champarna is today the district Champaran in northern Bihar bordering the Tarai district of Chitwan.

128. It has been very common among so-called "tribals", such as the Hor, the Sattar, the Hos and the Munda, to refer to oneself as 'man' (cf. McDonaugh 1984a).

129. Tove C. Kittelsen (1996) has recently submitted her MA thesis on the Rana Tharu wedding practices, and Harald Skar is in the process of writing a Rana Tharu monography.

130. Pali is an Indo-European (Aryan) language, and the oldest Buddhist scriptures were written in Pali. Pali is thus not the ancient language of Sri Lanka, but was spoken in and around Bengal.

131. At the back of Singh's book, there is a publisher's note by a Tharu stating: "I have gone through many books published in Nepal and abroad relating to the Tharus, but none threw light on the tribal origin of the Tharus. By publishing this book, he has made the Tharus come to know themselves and be known in the world".

132. In the English version, Singh writes in relation to the Rajasthan origin: "Now I categorically reject these theories about the Tharu origin. What I have found in history is altogether different. I do not say this on mere conjecture but on the basis of solid facts from ancient Indian history." (ibid.:2).

133. Singh defends his interpretation: "This interpretation of the word 'Tharu' is logical and in keeping with their traditional values and behaviour. The Tharus live a simple life and are renowned for their honesty. Sociologists and anthropologists bear out this truth about the Tharus..." (ibid.:4).

134. Buddhism, however, has also hierarchical aspects, and a society organised around Buddhist values, is not necessarily an egalitarian society (see e.g. Ortner 1989).

135. The Tamangs are a population group of Tibeto-Burmese origin who originally were settled in the central hills. The Newars are the original inhabitants of the Kathmandu valley and a people speaking a Tibeto-Burmese language.

136. King Maharana Pratab Singh was a Rajput king of the Sisodia clan who died a glorious death while fighting against the soldiers of the Mughal Emperor Akbar in 1567 AD. The Maharana (from Sanskrit: "the great Rana") is now celebrated as the upholder of Hindutva.

137. I was, however, never able to see anything which looked like an enaction of this in the Rana wedding ceremonies I observed during my fieldwork.

138. Hardwar, where the two rivers Ganges and Jamuna meet, is one of the holy places within Hinduism, and Hardwar is considered an auspicious place to die for Hindus. Hindus all over the world come here to write their genealogies with the Hindu scholars/priests (pandits) and each nationality/country has a special pandit. When I asked the pandits in Hardwar about a possible connection between the Rana Tharus and the Rajputs, they refused any such connection. There are so many people who claim a Rajput descent, they emphasised (in a private conversation, March 1994).

139. Hari Lal thought that this bulletin would be available at the National archives in Kathmandu. When I arrived there, I was sent to a special archive outside Kathmandu. But because the documents have not been micro-filmed, and are now in the process of falling apart, I was not allowed to look for this bulletin. There had, however, been several bulletins by that name, I was told. At the School of Oriental and Asian Studies (SOAS) in London, nobody had heard of any such bulletin.

140. It is not only Hari Lal who has reacted upon the way they have been described by outsiders. When I met one of the board members of the Tharu Welfare Society in Kathmandu, he was very upset by a recently published book about Tarai peoples. The author of this book, Hari Bansh Jha, argues that Tharus are not different from other Madhesi communities in the Tarai (Madhesi is the term generally used on people of the Tarai area, cf. Chapter 1) and Jha asks: "What purpose does the new word Tharu language serve?" The board member, however, emphasised that the Tharus were a people with a very rich culture who had lived in the Tarai since Buddha's days.

141. According to an historian at the City Palace in Udaipur, there were no Rajputs leaving Rajasthan during the Maharana's rule. Those who went to Nepal, Kashmir and Maharasthra, left during the first siège of Chittogarh, in the 14th century. During Maharana Pratab Singh, there was no johaar (collective suicide committed by the women). The famous johaar took place in 1534, whereas another took place in the fourteenth century.

142. Hasan uses the term Tharu when he talks about the Ranas of Kheri, and they are classified as Tharus in the official records. I do not know whether this classification "from above" coincides with their subjective identification. Or if Ranas in India, like their "brothers" in Nepal, insist that they are Ranas and not Tharus.

143. Although BJP is considered as a high caste party, BJP politicians and grassroots party workers try to mobilise lower castes around the party's Hindu nationalism (cf. G. Shah 1996).

144. Quoted by Charles Ramble in his article "Whither, Indeed the Tsampa Eaters" in Himal Sept./Oct. 1993: 21-25.

145. The term folklorisation, Jackson (1989) argues, is a negative term for describing processes of cultural self-consciousness, because it implies notions of inauthenticity. I will, however, for lack of another concept, use the term folklorisation, without suggesting that this has any inauthentic aspects.

146. This is not a new phenomenon. McDonaugh (1989) mentions a magazine called Gomcali, published by Tharus in Dang. This magazine, however, was deemed illegal during the panchayat system.

148. After what I have been told by some Indians from Bihar, this practice is common there as well. This may indicate that this is a regional practice, and not a particular "Tharu tradition".

149. This is pointed out by Keesing (1989:31) who writes: "Members of the Westernized elite are likely to be separated by gulfs of life experience and education from village communities where they have never lived: their ancestral cultures are symbols more than experienced realities". (See also Orans 1965:97).

150. Due to a bad infection, I was hospitalised in New Delhi and prevented from attending the culture programme.

151. One of the most prominent women in the pan-Tharu movement, told me that she used to be very unhappy with her cultural background. As the only female Tharu student at the Biratnagar campus, she hated Tharu culture, she told me in English. After becoming involved with the Tharu Culture Society, she had learnt to be proud of her culture.

152. Ram ram is a common way of greeting among friends in India, and it refers to the god Rama, the main person of the Ramayana epic already described in Chapter 1.

153. This party wants to have Hindi recognised as a national language. And it also claims right to citizenship for the many Madhesis who often do not have this.

Group	Number	Percentage
TARAI	5 718,770	30.9
a. Madhesi Caste groups	2 968,348	16.1
Baniya	101,868	0.6
Brahmin	162,886	0.9
Chamaar	203,919	1.1
Dhobi	76,594	0.5
Dusadh	93,242	0.5
Halwai	44,417	0.2
Kanu	70,634	0.4
Kayastha	53,545	0.3
Kewat	101,482	0.5
Khatwe	66,612	0.4
Kumahaar	72,008	0.4
Kurmi	166,718	0.9
Kusawah	205,797	1.1
Mallah	110,413	0.6
Musahaar	141,980	0.8
Rajbhat	33,433	0.2
Rajput	55,712	0.4
Sudi/Kalwaar	162,046	0.9
Teli	250,732	1.4
Yadav	765,137	4.1
b. Adivaasi groups	1 452,652	7.9
Dhanukh	136,944	0.7
Dhimal	16,781	0.1
Gangai	22,526	0.1
Rajbansi	82,177	0.4
Tharu	1 194,224	6.5
c. Other Madhesi groups	1 297,770	7.0
Bengali	7,909	0.0
Marwari	29,173	0.2
Muslim	653,055	3.5
Sikh	9,292	0.1
Others	627,514	3.4

Contents

Tharu Publications

Tharu Welfare Society

Magazines

British Gazetteers