silesr2014 009.

A Sociolinguistic Study of

Dangaura Tharu and

Related Varieties

Stephanie R. Eichentopf and

Jessica R. Mitchell

Dangaura Tharu and Related Varieties

Stephanie R. Eichentopf

and

Jessica R. Mitchell

SIL International

®

2014

SIL International Electronic Survey Report 2014-009, December 2014 ©2014 SIL International®

Abstract

This report presents the results of sociolinguistic research conducted among the Dangaura Tharu,

Kathariya Tharu and other closely related varieties in Dang, Banke, Bardiya and Kailali districts in Nepal. Lexical similarity, language attitudes, vitality and desires for language development were studied in order to assess how to best serve and support each of these communities, including which varieties are able and desire to use materials that are currently being developed in Dang district. Based on the findings of this research, it is suggested that Dangaura and Deukhuri varieties appear most likely to be usable and accepted by other varieties for material development. All of the communities we visited expressed desire for and appear ready to engage in developing multilingual education (MLE) program.

iii

Contents

Abstract Preface

1 Introduction

1.1 Geography— districts 1.2 The Tharu people

1.2.1 History 1.2.2 Culture 1.2.3 Education 1.3 Languages

1.3.1 Dangaura 1.3.2 Deukhuri 1.3.3 Malhoriya 1.3.4 Desauriya 1.3.5 Kathariya 1.4 Previous research

1.4.1 Previous linguistic research

1.4.2 Previous research on language vitality

2 Research questions

3 Methodology

3.1 Site selection 3.2 Research methods

3.2.1 Wordlists

3.2.2 Knowledgeable Insider Questionnaire (KIQ) 3.2.3 Informal interviews

3.2.4 Recorded Story Questions (RSQ) 3.2.5 Participatory Methods (PM) 3.3 Subject selection

3.3.1 Wordlists and RSQ elicitation subject selection 3.3.2 Informal interview subject selection

3.3.3 Recorded Story Question subject selection 3.3.4 Participatory Method subject selection

4 Lexical similarity

5 Dialect attitudes and emic perspectives

5.1 Attitudes and perspectives toward Dangaura 5.2 Attitudes and perspectives toward Deukhuri 5.3 Attitudes and perspectives toward Malhoriya 5.4 Summary of attitudes and emic perspectives

6 Language vitality

6.1 Domains of language use 6.2 Intergenerational transfer 6.3 Group dynamics

6.4 Network of social relations 6.5 Relative prestige

6.6 External recognition 6.7 Economic base 6.8 Vitality summary

7 Kathariya and Desauriya locations

7.1 Kathariya 7.2 Desauriya

8 Desires for development

Appendix A: Wordlists

Appendix B: Knowledgeable Insider Questionnaire (KIQ) Appendix C: Informal Interviews

Appendix D: Recorded Story Questions (RSQ)

Appendix E: Dialect Mapping

Appendix F: Appreciative Inquiry

Appendix G: Expanded Graded Intergenerational Disruption Scale References

v

Preface

This sociolinguistic survey of Dangaura Tharu, Kathariya Tharu and other closely related varieties in western Nepal was conducted in partnership with the Linguistic Survey of Nepal (LinSuN), Tribhuvan University, Kirtipur, Nepal. The data collection portion of this survey was carried out in September and October of 2011 in Dang, Banke, Bardiya and Kailali districts of Nepal. The purpose of the survey was to determine how best to serve each of these communities, including if Dangaura materials could be used by other groups.

Many people contributed to the completion of this language survey. Our colleagues at Tribhuvan University gave encouragement and support for which we are grateful. Background research, formation of tools, introductions to invaluable contacts and much of the fieldwork was completed under the guidance and direction of Edward Boehm, without whom this survey could not have been successful. We could not have completed fieldwork without working together with KC and BC, who traveled with us and arranged logistics, spent hours helping us translate and communicate what we were doing and were a profound asset to our team. We are also thankful for the guidance and support from our mentor, John Eppele, and colleague, Holly Hilty.

We are grateful to the Tharu people of the districts we visited, who welcomed and helped us. There are too many people to mention by name, but each of you had a part in this report coming to

completion. This report is for you.

It is our hope that this report will be a useful tool for Tharu language development. We trust that this report accurately reflects the data we collected. However, comments and suggestions are welcomed. July 2012

Stephanie R. Eichentopf Jessica R. Mitchell Kathmandu, Nepal

1

1 Introduction

This survey examines the relationships between Dangaura Tharu [thl] and its closely related varieties of Deukhuri, Desauriya and Malhoriya. It also seeks to examine Kathariya Tharu [tkt] in relation to these varieties. The purpose of this survey is to determine if Dangaura materials can be used by these other groups and, if not, identify which variety would be best for an adaptation. To better understand if literacy materials that are currently being developed by various non-governmental organizations (NGOs) in Dangaura Tharu may be usable by other varieties, an examination of lexical similarity, acceptability and vitality is needed. “The Dangora Tharu of Dang Valley should be further tested to the west and to the east of their respective centers to see what the limits are for intelligibility of each dialect. Little testing has been done in the eastern part of Kailali district and in the Bardiya and Banke districts of Nepal. These districts contain large and mixed populations of Tharus. The speech of these populations needs to be assessed to determine which dialect would most likely fit their literacy needs” (Hugoniot 1996:16).

For the purposes of this survey, the term Dangaura will be used to refer to only Dangaura Tharu speakers located within northern Dang district. All Dangaura related varieties in this survey will be identified by their variety name.

Another important distinction that should be made is the term Tharu. In published articles and books, the term Tharu sometimes refers to all types of Tharu people living along the Nepal-India border, including Chitwan Tharu, Rana Tharu, Dangaura Tharu, etc. However, the term sometimes refers to a specific Tharu group, such as Dangaura Tharu. For the purposes of this report, the term Tharu will be used only when discussing all Tharu varieties addressed in this report (Dangaura, Deukhuri, Desauriya, Malhoriya and Kathariya), unless otherwise noted.

1.1 Geography— districts

Nepal is currently divided into 14 administrative zones and subdivided into 75 districts. This survey spans across four districts in southwest Nepal. From east to west, these districts, in southwestern Nepal, are Dang, Banke, Bardiya and Kailali. They are within the Rapti, Bheri, Seti and Mahakali zones. The Mahendra highway runs east-west through the middle of all these districts (see map 1).

Map 1. Districts of Nepal

Source: UN map,

http://un.org.np/sites/default/files/2011-11-15-Nepal_Admin_Zone_Muni_A4_26August2011_v03.pdf (November 3, 2015)

Dang District

Dang district is widely recognized as the original homeland of Dangaura and Deukhuri Tharu. Dang district is sometimes identified as Dang-Deukhuri district. This is because in the past, it was divided into two districts: Dang to the north of the Mahendra highway and Deukhuri to the south of Dang Valley. The separation of Dangaura and Deukhuri speakers follows the geography of the district. The highest

concentration of Dangaura speakers is found in the wide Dang Valley in the north of the district,

bordered on the north by the Mahabharat Range and on the south by the lower Dang subrange. Deukhuri speakers are concentrated in the Deukhuri Valley, a long, narrow valley in the south of the district separated from Dang Valley by the Dang subrange and from India by the Dundwa subrange.

Map 2. Dang district

Source: Wikimedia CC BY-SA 3.0, https://en.wikipedia.org/wiki/Dang_District,_Nepal#/media/File: NepalDang Districtmap.png (November 3, 2015).

Banke District

Banke district contains a dense mixture of various Tharu varieties primarily located south of the Mahendra highway. Very few are found north of the highway or in the hill area of the southeast. Desauriya speakers are found in the southwest and west of Nepalgunj.

Map 3. Banke district

Source: Wikimedia CC BY-SA 3.0, https://en.wikipedia.org/wiki/Banke_District#/media/File:Nepal BankeDistrictmap.png (November 3, 2015).

Bardiya District

We did not visit any villages in Bardiya district during our fieldwork. However, many Tharu speakers are found there and these speakers are mentioned by interview responses and discussions. Different spellings can be found for the name of this district, the most common being Bardia and Bardiya. South of

Mahendra highway there are high concentrations of Deukhuri and Dangaura speakers, however, their exact locations are unknown. Malhoriya speakers are located in western Bardiya between the

bifurcations of the Karmali River. Few Tharus are located north of Mahendra highway because of the Royal Bardiya Reserve located there.

Map 4. Bardiya district

Source: Wikimedia CC BY-SA 3.0, https://en.wikipedia.org/wiki/Bardiya_District#/media/File:Nepal BardiyaDistrictmap.png (November 3, 2015).

Kailali District

One estimate suggests that as high as forty percent of the Dangaura Tharu population resides in Kailali district (Edward Boehm p.c., 2011). Within Kailali, Malhoriya speakers are found concentrated around Tikapur in the southeast. Many Tharu speakers are found in the southwest of the district. Few Tharu speakers are located north of Mahendra highway.

Map 5. Kailali district

Source: Wikimedia CC BY-SA 3.0, https://en.wikipedia.org/wiki/Kailali_District#/media/File:Nepal KailaliDistrictmap.png (November 3, 2015).

1.2 The Tharu people

The Tharu ethnic identity is comprised of many separate groups that vary in culture and language. Generally, the term Tharu refers to a large ethnic group that spans most of the Tarai in Nepal. In

actuality, there are many diverse ethnic and language groups under the umbrella term Tharu. “Outsiders generally view the Tharu as one homogeneous group… The Tharu, however, recognize many different subgroups distinguished by clan, region, cultural differences, and language” (Webster 1993:4). The most well-known of these groups are the Rana, Dangaura, Kochila/Morangia, Chitwania and Kathariya (Boehm 1997:19).

1.2.1 History

Dangaura speakers are originally from Dang district, but currently live not only in Dang, but also in the districts west of Dang. This westward migration occurred as a result of a push for development of the Tarai in the 1950s. Prior to this time, the Tharu were one of the few people groups with an immunity to malaria, allowing them to live in relative isolation. As the Malaria Eradication Project lessened the threat of malaria and roads were constructed, hill people were encouraged to settle and work the land in the Tarai. These newcomers had more political awareness and gained control of the land in regions like Dang (Guneratne 2002:4, 94). In response, many Dangaura Tharu speakers migrated west (van Driem 2001:1167).

1.2.2 Culture

The ethnic identities of the Tharu people groups in this survey do not necessarily align with their linguistic identity. Different Tharu groups maintain their own set of religious rituals, marriage patterns and family relations. This section summarizes the economy, religion and family structure of Tharu varieties of this survey.

According to observations of the Webster 1993 survey, the western Tharu were almost completely rural, living in villages of less than 2,000 people. Their primary economic base was agriculture and they were generally self-sufficient year round through hunting, fishing and gathering (Webster 1993:3, 113). Referring to the western Tarai, Rajaure says, “a Tharu village consists of several families living in a compact social unit which benefits them both socially and economically…” (1981:155). Marriage tends to occur within the same clan, but not in the same village. According to McDonaugh and confirmed through observations and interviews, Rana, Kathariya and Dangaura Tharu people do not intermarry but think of each other as separate castes (1989:193). Women go to live with their husband’s family. Of Dangaura Tharu within Dang district, Rajaure reports that they rarely leave Tharu areas for more than a couple of days (1981:178). These reports were verified through informal interviews where 78% of respondents say they have never travelled outside of their area.

Most Tharu people practice a traditional religion, however many claim to be Hindu (Boehm 1997:27). “Most Tharu practice an indigenous form of animism, in which shamanism, ancestor worship and tattooing play pivotal roles” (van Driem 2001:1167). Idol and artifact worship is common.

Guneratne points out that following other groups migrating into the Tharu area after the malaria eradication in the 1950’s, “Tharus were no longer able to perform customary rituals, because they no longer had the resources to do so; these rituals became abbreviated in their performance or fell into disuse entirely” (1998:760).

1.2.3 Education

Education is increasingly available in many Tharu areas. Village leaders report that “half” to “most” of their communities are literate. We observed that while the majority of men are literate, women are predominately illiterate, though this seems to be changing with the younger generation. The Backward Society Education NGO is one of the most influential advocates for Tharu education.

1.3 Languages

The focus of this study is on Dangaura Tharu [thl] and other closely related varieties, including

Deukhuri, Malhoriya and Desauriya. In addition, it also looks at Kathariya Tharu [tkt]. The 1991 census figures for Tharu populations were reported by district, but do not include population figures for each individual variety. According to the census data, Dang district has a Tharu population of 354,413 (29.6% Tharu), but again, it is unknown how many are Dangaura Tharu and how many are Deukhuri Tharu, as both reside in Dang district. Tharu population in Banke district is 285,604 (15.6% Tharu), 290,313 in Bardiya (51% Tharu) and 417,891 (47% Tharu) in Kailali. This section describes each variety in more detail.

1.3.1 Dangaura

Dangaura Tharu is an Indo-European, Indo-Iranian, Indo-Aryan, Central zone language written in Devanagari script. Other names and spellings for the language are Chaudary Tharu, Chauhari Tharu, Dangauli, Dangha, Dangora and Dangura. Dangaura, in which materials are currently being developed, is primarily spoken in Dang district; however, it is also found in the districts to its west as well as in Kheri, Gonda and Bahraich districts of India (Boehm 1997:19). Nepali has influenced this speech variety.

1.3.2 Deukhuri

Southern Dang district is viewed as one of the original homelands of Deukhuri speakers. They are also located in districts to the west. Deukhuri is listed in the Ethnologue with the same ISO code as Dangaura [thl]. While allowing intermarriage between their groups, Dangaura and Deukhuri speakers identify themselves as separate. Because Deukhuri communities are so close to India, Deukhuri has been influenced by Hindi. The Curriculum Development Centre (CDC) of the Ministry of Education of the Government of Nepal has done some development work in this variety.

1.3.3 Malhoriya

Malhoriya is located in eastern Kailali district around Tikapur, as well as in western Bardiya in the original location of the Dangaura westward migration. Malhoriya shares the same ISO code in the Ethnologue as Dangaura [thl].

1.3.4 Desauriya

Desauriya, sometimes spelled Deshauriya, is not included in the Ethnologue. “The Desauriya Tharu are a related group [to Dangaura Tharu] also inhabiting Banke and Bardiya districts” (van Driem 2001:1167). The origin of Desauriya speakers is unknown, but it is generally believed that that they did not originate in the Dang/Deukhuri area. “In Bardiya and Banke districts, the Dangaura Tharus distinguish themselves from the Deshaurya Tharus. But Deshauriya culture is so close to that of Dangaura that I presume they could be an offshoot of an earlier wave of migrants from Dang…” (Krauskopff 1995:187–188).

Additional information about the location of Desauriya speakers is in section 7.2.

1.3.5 Kathariya

Kathariya Tharu is listed in the Ethnologue under ISO [tkt]. An alternate name is Khatima Tharu. Often, it is spelled Kathoriya. However, a leader of their community in Pabera, Kailali indicated to us that the correct spelling is Kathariya. It is classified as Indo-European, Indo-Iranian, Indo-Aryan, Central zone, unclassified. “Kathoriya Tharu are concentrated most heavily in Kailali district of Nepal but can also be found in Kheri, Bahraich, Gonda, and Gorakpur districts in India and in other areas where Dangaura Tharu are found” (Boehm 1997:19). Hugoniot’s research claimed that all Kathariya villages have Dangaura speakers living in them. In addition, he reported that only Tharus living near Kathariya villages have ever heard of Kathariya Tharu (1996:3, 13). The interviews and observations of this survey verify Hugoniot’s claims that most Tharu (specifically Dangaura, Deukhuri, Desauriya and Malhoriya) have neither heard of nor interact with Kathariya. The origin of the Kathariya is unknown, but it is generally believed that, unlike other varieties included in this survey, they did not originate in the Dang/Deukhuri area. Additional information about the location of Kathariya speakers is in section 7.1.

1.4 Previous research

There has been a considerable amount of research on the economics, religion and customs of the Tharu community, but relatively little linguistic research has been published.

1.4.1 Previous linguistic research

“The only significant studies on the language of the Tharus include the work by Frierson (1903) and Hodgson (1857), the atlas of Bhojpuri by Tivari (1960), an inventory of phonemes in Citvan Tharu by D. Leal (1972) and Citvan Tharu sentences by William Leal (1973)” (van Driem 2001: 1166). Previous research by Boehm (1997), Hugonoit (1996) and Webster (1993) has provided some information about Kathariya and Dangaura, although their research primarily focused on other Tharu varieties. Little or no

research is found regarding the language varieties of this survey. The most detailed published research on the groups of this survey are the theses: “A Phonological Reconstruction of Proto-Tharu and Language Use” (Boehm 1998) and “Language Maintenance Among the Tharu of the Indo-Nepal Tarai” (Boehm 1997).

1.4.2 Previous research on language vitality

Tharu speakers’ degree of exposure to and contact with other languages (specifically Nepali and Hindi), makes language vitality necessary to investigate. Kelly Boehm’s thesis (1997) addressed this question. Several factors surfaced as pointing towards healthy language maintenance:

• Government attitudes toward minority languages and cultures—Given the current political stance toward preserving minority languages and cultures, many groups (including Tharu) have received encouragement from the government for language and culture preservation.

• Attitudes toward and practices of exogamy—Through questionnaires, Boehm found that 70% of Tharus did not approve of marrying a non-Tharu. Tharu groups differ greatly, which means that many will not marry outside of their specific group.

• Positive language attitudes—Boehm’s questionnaire revealed that 91% of Tharu speakers believed their language to be better than the language of wider communication (Nepali or Hindi), 96% thought that mothers should speak Tharu to their children and 98% anticipated that Tharu would continue to be spoken in the future.

• Education—At the time of Boehm’s study, education was not accessible enough for the Tharu to affect the vitality of the language. She recommended that this issue be re-evaluated when it becomes a possible factor in language maintenance.

2 Research questions

The purpose of this survey is to see how many Tharu varieties can use materials that are being developed in Dangaura Tharu and to investigate how to best serve the speakers of Deukhuri, Malhoriya, Desauriya and Kathariya.

The five goals of this survey are:

1. Lexical Similarity: Investigate lexical similarity between all Tharu varieties of this survey.

2. Dialect Attitude Assessment: Assess the attitudes of the language varieties of this survey toward one another to better understand their willingness to share written materials.

3. Language Vitality: Evaluate the vitality of Deukhuri, Desauriya, Malhoriya and Kathariya varieties. 4. Kathariya and Desauriya Locations: Identify the major population centers of Kathariya and

Desauriya speakers.

5. Desires for Development: Discover each community’s desires for their own language-based development.

3 Methodology

3.1 Site selection

A total of seven villages were visited in three districts. We gathered data in one location for each variety, except the Deukhuri variety, which had three sites. Site selection was based upon where the variety is known to be spoken and the recommendation or availability of contacts.

Deukhuri and Desauriya sites were located on or near the Mahendra highway, a factor which raised concerns regarding the probability of high contact with other languages. High contact with other

languages can affect language vitality. The high levels of vitality we encountered, however, lead us to believe that data gathered at these sites are a strong argument for language vitality, since language vitality generally increases with distance from population centers. It is probable that language vitality would be even stronger in more remote villages.

During site selection, we were unaware that Chapargaudi, the Desauriya village we visited, has only a 10% Tharu population and is highly mixed. Many of our subjects were from nearby villages ranging from 1–3 km in distance from Chapargaudi; however, they view themselves as one group with no differences in their language.

3.2 Research methods

Several methods were used to elicit data in order to meet the goals of the survey. Sections 3.2.1 through 3.2.5 describe the purpose, procedure and certain advantages and disadvantages of each method.

3.2.1 Wordlists

Description and purpose: A comparison of wordlists estimates the degree of lexical similarity between the speech varieties the wordlists represent.

Procedure: Five 316-item wordlists were elicited and compared for this analysis, one from each Tharu variety. Wordlists were elicited in Nepali from Tharu mother-tongue speakers and were transcribed using the International Phonetic Alphabet (IPA). Lexical similarity analysis was carried out on each pair of wordlists. A complete description of wordlist comparison methodology and the data we collected can be found in Appendix A.

Advantages: Data collection is relatively efficient. Wordlists can provide some broad insights into possible dialect groupings.

Disadvantages: Above certain levels of lexical similarity, wordlists cannot give conclusive evidence of intelligibility between speech varieties compared.

3.2.2 Knowledgeable Insider Questionnaire (KIQ)

Description and purpose: A prepared interview specifically designed for someone the community views as the most knowledgeable regarding information about their language. This tool provides information from a reliable and knowledgeable source about their language and social factors that do not vary from person to person.

Procedure: Administered to only one person at a time, questions range from specific population estimates and locations to general information about vitality and other languages spoken by the community. This tool was administered in one village from each speech variety of the survey. The questionnaire and responses are found in Appendix B.

Advantages: Good for obtaining village-level information in a brief period of time. Disadvantages: Information is from only one person and therefore may be skewed.

3.2.3 Informal interviews

Description and purpose: A prepared interview schedule (based on the “Sociolinguistic Questionnaire A”, used by the Linguistic Survey of Nepal) guides interaction in order to gather information regarding specific sociolinguistic issues, while allowing freedom to ask further questions if they might provide additional information relevant to the research questions of the survey.

Procedure: Interviews were administered on an individual basis. Subjects were chosen using quotas according to demographic factors relevant to our research questions (see section 3.3.2 for additional subject selection procedures). This tool was administered to speakers of all the Tharu varieties of this survey except Dangaura1. _{The complete interview schedule and responses are in Appendix C.}

Advantages: Depending on the length of the interview, the time in administration can be minimal, allowing for relatively large numbers of people to be interviewed. The informal nature of the interviews helps subjects feel comfortable to share openly, while allowing greater depth and providing context for their responses.

Disadvantages: Informal interviews are limited in that subjects may only report what they want the researcher to hear, or what they believe the researcher would like to hear.

3.2.4 Recorded Story Questions (RSQ)

Description and purpose: Subjects listen to recorded stories, with questions asked after the stories. This helps in the assessment of the subjects’ perceived understanding of and attitudes toward actual samples of the language from various areas. It helps the researcher understand if subjects perceive different speech samples as representing their own style of speech.

Procedure: A narrative story is recorded in one location and played for subjects in a second location. The collected story is first tested with at least ten mother-tongue speakers of the variety before administering the test in other locations. The subjects in the second location are not told where the storyteller is from. After listening to a story, subjects answer questions about their understanding of and opinions toward the storyteller’s speech. The RSQ, RSQ responses and translated stories are all found in Appendix D. Advantages: Evaluates perceived understanding of opinions toward actual samples of the language. Disadvantages: Decisions about dialect areas should not be based solely upon these responses and are best used in conjunction with more complete intelligibility testing.

3.2.5 Participatory Methods (PM)

This survey used two different participatory methods (PM): Dialect Mapping and Appreciative Inquiry. The purpose for using PM is to gain perspective from the community regarding what they see happening with their language. PM attempt to understand the emic perspectives of the community.

Dialect mapping

Description and purpose: This tool creates space for discussion of emic perspectives regarding dialects, their geographic location and perceived levels of comprehension between varieties.

Procedure: Participants are invited to describe their linguistic landscape by identifying other locations where their language is spoken. They then identify how large they perceive the differences to be between their variety and the others as well as their level of understanding, which variety they use in

conversation with people from other areas and which variety they believe to be the standard or most broadly understood. Two trained Dangaura Tharu speakers facilitated dialect mapping in five locations. Full steps used for this facilitation are in Appendix E.

1 _{This is because material development is already in progress in Dangaura and therefore less information was needed} from Dangaura speakers than from other varieties in order to fulfill the purpose and goals of this survey.

Advantages: Provides a visual representation of other communities participants interact with, how well they feel they understand those varieties, how their language may or may not be altered in these circumstances and their attitudes about other varieties.

Disadvantages: May seem complicated or redundant, but each step contributes to a fuller picture of the local perception of the language situation. Emic perspectives do not always match linguistic reality. Appreciative Inquiry

Description and purpose: This tool helps community members dream and discuss the possibilities for their language and begin planning how they can achieve those dreams. It also shows what the community sees as most important for their language.

Procedure: Participants discuss things in their language or culture that have made them happy or proud. They are then invited to think about how they might build upon these good things they have identified, or identify dreams they have for their language. Next they discuss what dreams might be accomplished sooner and which ones will take longer. Then they identify which dreams are most important to them. Finally, participants are asked if they would like to choose a dream they would like to create a plan for and begin deciding on what the first steps should be, who will be involved and when the plan will be put into action. Like dialect mapping, this was facilitated by two trained Dangaura Tharu speakers. Complete steps of Appreciative Inquiry are in Appendix F.

Advantages: Creates space and opportunity for community discussion of good things that are currently taking place, their goals and dreams and concludes with creating their own plan of action for a goal they have chosen (if they so wish). It is adaptable for various situations.

Disadvantages: It requires facilitators to have strong linguistic and cultural command for effective group facilitation. If not properly framed, the facilitation may raise false hopes of outside assistance in reaching their goals.

3.3 Subject selection

Subject quotas for this survey are based on a convenience sample. It focuses on four demographic groups (gender, age, education and language variety), as these factors are known to influence language use and attitudes. People in these demographic groups often have varying levels of exposure to other languages.

3.3.1 Wordlists and RSQ elicitation subject selection

There are four requirements for subjects chosen for wordlists and recorded story elicitation. These criteria are as follows:

1. Subject has grown up in the village under study, lives there now and if they have lived elsewhere, it was not for a significant amount of recent time2_.

2. Subject has at least one parent from the target mother tongue.

3. Subject has at least one parent from the village under study and that parent spoke the mother tongue (L1) with them.

4. Subject speaks L1 first and best.

2 _{It is difficult to define a specific time period (e.g. more than the last five years) for “a significant amount of recent} time.” Thus, this criterion is intentionally subjective as it depends on how long the subject lived elsewhere and how long they have been back in the village relative to their age.

3.3.2 Informal interview subject selection

The informal interview schedule requires that only criteria numbers one and two (section 3.3.1) be met in order for a subject to be eligible. In each language variety a minimum of 12 informal interviews were administered stratified by age and gender. Educational background was also accounted for during data collection and analysis. For this analysis, literate persons were classified as educated, which generally corresponded with the completion of primary level four.

Table 1. Sample size for informal interview by age and gender

Sample size by strata Age Total

Young (15–34) Old (35+)

Gender Male 3 3 6

Female 3 3 6

Total 6 6 12

3.3.3 Recorded Story Question subject selection

Subjects chosen for listening to and responding to the RSQs met the first three screening criteria that was used for the RSQ storyteller (section 3.3.1). In each language variety a minimum of ten RSQ tests were administered.

3.3.4 Participatory Method subject selection

Appreciative Inquiry and dialect mapping are facilitated for groups in each community. There is no limit imposed on how many people may be involved although a group size between five and ten individuals is preferable. It is also preferred for various demographics (gender, age and education) to be represented in each group if possible. There is no screening process for those involved in participatory methods.

However, notes were taken regarding who was present in order to account for possible bias.

4 Lexical similarity

Lexical similarity is a measure of the relative similarity of a sample of words from two speech varieties. Similarity percentages are determined by calculating the percentage of words in one speech variety that are pronounced the same or in a very similar way to the words in another speech variety. Specifically, this survey used the comparison method outlined in (Blair 1990: 31–32), further explained in Appendix A.2. It is generally accepted that lexical similarity percentages below 60% indicate that the compared lists represent different languages. Lexical similarity above 60% requires intelligibility testing to confirm if the varieties are dialects of the same language or if they are different languages. This survey compares five 316-item wordlists. The lexical similarity percentages among the five Tharu wordlists are displayed in table 2.

Table 2. Lexical similarity percentages matrix

Malhoriya

88% Deukhuri

83% 84% Dangaura

84% 81% 76% Desauriya

Table 2 shows that the lexical similarity percentages vary from a high of 88% (between Malhoriya and Deukhuri) to a low of 70% (between Dangaura and Kathariya). No comparison falls below the cutoff (60%) where they would be considered separate languages. Malhoriya and Deukhuri have the highest average lexical similarity with all other varieties.

The wordlist from Dangaura shared higher similarity with those taken in Malhoriya (83%) and Deukhuri (84%) than with Desauriya (76%) or Kathariya (70%). This suggests a greater potential for intelligibility challenges with Desauriya and Kathariya.

Wordlist data shows that Malhoriya and Deukhuri share higher lexical similarity with Dangaura and Desauriya. Dangaura shares high lexical similarity with Deukhuri and Malhoriya. Kathariya is the most lexically variant of these wordlists.

5 Dialect attitudes and emic perspectives

One goal of this survey was to investigate the emic (insider) perspectives of each community in regards to dialect boundaries and assess attitudes toward Dangaura, Deukhuri and Malhoriya varieties. This gives a more complete overview of the language situation than linguistic data alone can provide. Although reported similarity requires further testing, reported differences are generally more reliable. The

following sections are divided into each variety (Dangaura, Deukhuri and Malhoriya) and begin with the opinions expressed by dialect mapping participants regarding their own language situation, including perceived boundaries in speech differences. Informal interview responses are then discussed which suggest minimal contact between varieties and opinions regarding different speech varieties. Lastly, attitudes toward the variety are discussed, based on data collected through RSQs and other tools (when applicable). Although the interview and RSQ sample size in each site was not large, we believe they still provide helpful indicators towards these research questions.

One of the questions included in dialect mapping was about the development of written materials in Tharu. When groups were asked, assuming no materials had yet been developed, what the best variety for material development would be, each chose their own variety as their first choice. The next nearest (geographically) variety was always their second choice. This provides no consensus on any one dialect that all groups agreed would be best for the development of materials in Tharu.

Given the large geographic area covered by the varieties in this survey (nearly 200 km between the farthest two villages), it is not surprising that speakers of some speech varieties have never heard of nor had contact with speakers of some other varieties. During dialect mapping facilitations in this survey, this was often the case. For example, participants in Belganar, Dang (Dangaura Tharu) had heard of Malhoriya and Kathariya varieties, but had never met anyone that speaks either of those varieties.

5.1 Attitudes and perspectives toward Dangaura

This section summarizes the data collected regarding attitudes and perspectives toward Dangaura. Through a dialect mapping facilitation in Belganar, Dang, Dangaura participants shared how they view the boundaries of their language and its varieties. They listed many Village Development

Committees (VDCs) that they completely understand, all of which lie north of the Mahendra highway: Narayanapur, Bijouri, Manapur, Duruwa, Phulbari, Goltakuri, Hekuli, Shreegaun and Ghorahi. They reported that they understand half of the speech in Satbariya VDC, which lies just south of the Mahendra highway in what was once called Deukhuri district. When asked about the other Tharu varieties of this survey, participants reported that they understand half of Deukhuri and Desauriya speech. They did not have knowledge about Malhoriya or Kathariya and those groups were excluded from the facilitation.

Participants reported that Deukhuri speakers adjust to them by speaking the Dangaura variety, although Deukhuri participants did not report this about themselves.

Informal interviews were used to inquire about minimal levels of contact between speech varieties and the opinions of speakers of these different varieties toward each other. As seen in table 3, tables throughout this report include a row for the total number of responses indicated by N=. The total number of responses vary for different questions because some questions were not applicable to some

subjects. In addition, abbreviations for each Tharu variety are displayed in tables throughout this report as follows: Dangaura (DG), Deukhuri (DK), Desauriya (DES), Malhoriya (MAL) and Kathariya (KAT). Table 3 displays the number of respondents from each variety that reported meeting speakers of Dangaura.

Table 3. Contact with Dangaura speakers

DK MAL DES KAT Total

N= 12 13 12 12 49

Have met DG

11 (92%)

13 (100%)

11 (92%)

12 (100%)

47 (96%)

Table 3 shows that the majority of respondents from all varieties report contact with Dangaura speakers (47/49). This is not surprising because Dangaura speakers comprise the most populous group and live in villages with or near speakers of other varieties in every district.

During informal interviews, if respondents reported meeting speakers of the different speech varieties, they were also asked how they liked their language. Responses to this question by those that have met Dangaura speakers are shown in table 4.

Table 4. Informal interview responses regarding Dangaura speech

DK MAL DES KAT

N= 11 13 12 12

How did their [DG] Tharu make you feel?

Good 82% 100% 100% 67% Indifferent 18% – – 33%

Bad – – – –

Most participants who said they have met Dangaura speakers reported that they thought the Dangaura speech of those they met is “good.” Kathariya respondents reported mixed attitudes: the majority (67%) felt the Dangaura speech was “good” and 33% expressed indifference. None of the respondents reported the speech as “bad.” This shows lack of negative attitudes by those who have come in contact with Dangaura speakers.

Another tool administered to assess attitudes was the RSQ. After listening to the Dangaura story, listeners were asked to share what they thought of the storyteller’s speech and whether or not the speech is different from their own. Results of these questions are shown in table 5.

Table 5. Attitudes towards speech in Dangaura story

DK MAL DES KAT

N= 10 13 12 11

How did you like his speech?

Good 90% 77% 92% 64%

Fine 10% 23% 8% 36%

Bad – – – –

Is the language in this story…from the language

spoken here?

Same 40% – 25% –

A little different 60% 69% 67% 45% Very different – 31% 8% 55%

When listening to an actual sample of Dangaura speech, most respondents said the language variety they heard was “good.” None of the respondents from any variety reported the speech as “bad.” This is consistent with informal interview responses from those that had met Dangaura speakers, further suggesting positive attitudes toward Dangaura.

Table 5 shows that all Kathariya and Malhoriya respondents reported the Dangaura story was different in some way from their own speech. Most (75%) Desauriya respondents also reported differences between Dangaura and their own variety. Despite this, the majority of these respondents reported the speech as “good.” None of the respondents from any area responded with clearly negative attitudes towards Dangaura. The fact that these respondents identify the Dangaura storyteller’s speech as different from their own and still report that the speech is “good” suggests generally positive attitudes toward Dangaura speech. When the story was identified as “different,” participants were asked to identify what the differences were: style, pronunciation and/or vocabulary. Style and pronunciation were identified by respondents as the top factors making the Dangaura speech different from their own.

When asked if they would allow their son or daughter to marry someone who spoke like the storyteller, an average of 89% said “yes,” suggesting negative attitudes are not held toward Dangaura speech. Kathariya respondents reported the lowest percentage (73%).

After listening to the Dangaura story, 55% of listeners (25 of 46) were able to identify the

storyteller’s speech as Dangaura. Respondents were asked, “How much of the story did you understand: all, most, half, or less than half?” Table 6 displays responses to this question by subjects from each of the varieties.

Table 6. Perceived comprehension of Dangaura story

DK MAL DES KAT

N= 10 13 12 11

How much of the Dangaura story did you

understand?

All 100% 85% 83% 73%

Most – 15% 17% 18%

Half – – – 9%

Less than half – – – –

The pattern of reported comprehension follows a geographic trend, with each respondent in

Deukhuri reporting that they understood all of the story, followed by Malhoriya and Desauriya (85% and 83% respectively) saying they understood all of the story and finally 73% in Kathariya reporting full comprehension. This is consistent with dialect mapping responses from Deukhuri, Desauriya and

Malhoriya participants. The dialect mapping group in Deukhuri said that they understand most Dangaura speech. Deukhuri and Malhoriya participants said that they understand Dangaura best, after their own variety.

During dialect mapping, Desauriya participants reported that they use the Dangaura variety with Dangaura people, in order to improve communication. Their willingness to change their own speech with Dangaura speakers suggests that they view the Dangaura speech favorably. However, Desauriya participants also said they understand only half of Dangaura speech, compared to most (82%) reporting understanding “all” of the Dangaura story.

Through informal conversations, several community members reported the Dangaura speech variety as the most polite variety of Tharu.

In summary, the data shows that Dangaura is readily identified by other Tharu communities of the area and many report it is easy to understand. Even though other varieties identified the Dangaura speech as different from their own, the data suggests positive attitudes are held toward Dangaura speech. Dialect mapping exercises and observations support this as well. No interview responses or observations suggest there are negative attitudes held toward Dangaura speech. This is further confirmed through a high percentage of respondents reporting that they would allow intermarriage with Dangaura speakers.

Kathariya respondents reported more differences than the other varieties between their own language and Dangaura. Despite this, all but one respondent felt they understood most or all of the Dangaura story. Although this does not confirm comprehension of Dangaura by the Kathariya respondents, it shows that they have positive attitudes towards the Dangaura variety.

5.2 Attitudes and perspectives toward Deukhuri

This section summarizes the data collected regarding attitudes and perspectives toward Deukhuri. Dialect mapping participants in Majgau, Dang district, a Deukhuri village, listed several VDCs as having speech that sounds exactly like their own: Sonpur, Gobardiya, Gangapraspur, Gadhawa, Chaulahi, Lalmatiya and Sisahaniya. These districts are all located in western Dang district, south of the Mahendra highway, except Lalmatiya and Sisahaniya VDCs which are located north of the highway. Like Dangaura participants, they said they understand most, but not all off the Tharu spoken in Satbariya VDC. They also reported understanding most of the speech in Bela VDC. When asked, participants said they do not understand very much Desauriya. They had not interacted frequently enough with Malhoriya or Kathariya speakers to discuss those varieties during the facilitation. When asked which speech varieties they understand best, participants listed the same VDCs as they listed when asked who they completely understand. They reported understanding Satbariya VDC as second most understood, followed by Nepali.

Table 7 shows the minimal levels of contact between Malhoriya/Desauriya/Kathariya and Deukhuri reported in individual interviews.

Table 7. Contact with Deukhuri speakers

MAL DES KAT Total

N= 13 12 12 37

Have met DK

10 (77%)

6 (50%)

10 (83%)

26 (70%)

Most Malhoriya (10 of 13) and Kathariya (10 of 12) respondents have met Deukhuri speakers. However, only six (6 of 12) Desauriya people interviewed had contact with Deukhuri. Of those who reported meeting Deukhuri speakers, a follow-up question was asked: “How did their Tharu make you feel?” Responses are shown in table 8.

Table 8. Informal interview responses regarding Deukhuri speech

MAL DES KAT

N= 10 6 10

How did their [DK] Tharu make you feel?

Good 80% 50% 90%

Indifferent 20% 50% 10%

Bad – – –

When asked how they like the language of the Deukhuri people they have met, none of the respondents said the Deukhuri language was “bad.” This shows a lack of clearly negative attitudes toward Deukhuri speech by those who have met them.

Table 9 shows responses to RSQ questions asking how listeners liked the storyteller’s speech and if the speech was different from their own.

Table 9. Attitudes towards speech in Deukhuri story

DG MAL DES KAT

N= 10 13 12 11

How did you like his speech?

Good 80% 77% 92% 91%

Fine 20% 23% 8% 9%

Bad – – – –

Is the language in this story…from the language

spoken here?

Same 10% 69% 33% 20%

A little different 70% 31% 58% 40% Very different 20% – 8% 40% As seen in table 9, the majority of respondents said that the storyteller’s speech was “good.” Over 90% of both Desauriya and Kathariya respondents said that the storyteller’s speech was “good.” None of the respondents said they disliked the Deukhuri speech they had heard. This data suggests positive attitudes toward Deukhuri speech from the tested varieties.

The majority of Malhoriya respondents (69%) reported the Deukhuri storyteller’s speech as the “same” as their own. By contrast, only 33% of Desauriya, 20% of Kathariya and 10% of Dangaura respondents said the same about the storyteller’s speech (table 9). Regardless of whether respondents claimed the storyteller’s speech as the “same” as their own, the majority reported the speech as “good.” This suggests positive attitudes toward Deukhuri speech because, even when differences were

recognized, the speech sample was viewed as “good.”

The majority of respondents from all varieties reported that they would allow their son or daughter to marry someone who spoke like the Deukhuri storyteller. This suggests that the participants do not hold negative attitudes toward Deukhuri speech.

After listening to the Deukhuri story, participants were asked, “How much of the story did you understand?” Table 10 displays the responses.

Table 10. Perceived comprehension of Deukhuri story

DG MAL DES KAT

N= 10 13 12 11

How much of the Deukhuri story did you

understand?

All 100% 100% 92% 91%

Most – – 8% 9%

Half – – – –

Less than half – – – –

While fewer respondents have had previous contact with Deukhuri speakers than with Dangaura, table 10 shows that more respondents felt they understood all of the Deukhuri story than said the same about the Dangaura story. Consistent with this, Kathariya dialect mapping participants reported that they understand Deukhuri speech better than Dangaura speech. After Kathariya participants heard the Dangaura story, eight of 11 (73%) reported understanding all of the story compared to 10 of 11 (91%) claiming they understood all of the Deukhuri story. Every Dangaura and Malhoriya RSQ respondent reported understanding all of the Deukhuri story. However, Dangaura and Malhoriya dialect mapping

participants reported that they understand half of what Deukhuri speakers say. This indicates that Dangaura and Malhoriya speakers identify their speech as different from Deukhuri, but respond differently to an actual sample of speech.

Of those who reported the storyteller’s speech as different in some way, the primary reason was “style.”

In summary of what speakers of other varieties reported about Deukhuri speech, the Deukhuri story was reportedly the most understood story of all the varieties tested. A high percentage of respondents said the Deukhuri speech in the story was “good.” The same was reported through informal interviews. Desauriya and Kathariya respondents reported the highest amount of differences between their speech and Deukhuri. Despite identifying differences, the majority of respondents from all varieties reported favorable attitudes towards Deukhuri speech. In addition, the majority of interview respondents said they would allow marriages with Deukhuri speakers. Based on this data, it can be generalized that respondents do not hold negative attitudes toward Deukhuri.

5.3 Attitudes and perspectives toward Malhoriya

This section summarizes the data collected regarding attitudes and perspectives toward Malhoriya. Malhoriya dialect mapping participants in Ghiya, Kailali district, listed several VDCs located in southeast Kailali district as having speech that sounds exactly like their own: Chauha, Kota Tulsipur, Baliya, Janakinagar, Joshipur, Pathariya, Munuwa, Durgauli, Thapapur, Tikapur, Narayanpur and Dansinhapur. Participants reported that they understand Dangaura best, after their own variety. They reported understanding only half of both Deukhuri and Kathariya speech. During dialect mapping with other varieties, Malhoriya was rarely mentioned by participants. Therefore, this section will focus on responses about Malhoriya from RSQ and informal interviews.

Table 11 shows the minimal levels of contact between Deukhuri/Desauriya/Kathariya and Malhoriya reported in individual interviews.

Table 11. Contact with Malhoriya speakers

DK DES KAT Total

N= 12 12 12 36

Have met MAL

1 (8%)

2 (17%)

12 (100%)

15 (42%)

As seen in table 11, very few Deukhuri or Desauriya informal interview respondents said that they have met a Malhoriya speaker. Therefore, minimal data is available from informal interviews regarding attitudes toward Malhoriya speakers by those that have met them. All Kathariya respondents said they have met Malhoriya speakers. Only two Desauriya respondents and one Deukhuri respondent reported having met Malhoriya speakers.

Respondents that have met Malhoriya speakers were asked how they liked their speech. Kathariya responses to this question are shown in table 12.

Table 12. Informal interview responses regarding Malhoriya speech

KAT

N= 12

How did their [MAL] Tharu make you feel?

Good 50%

Indifferent 33%

Table 12 shows that half of the Kathariya respondents felt that the speech of Malhoriya people is “good.” Two Kathariya respondents did not like the Malhoriya speech that they heard in the story. This data shows mixed attitudes toward Malhoriya speech by Kathariya respondents.

The Malhoriya story was played for Deukhuri, Desauriya and Kathariya participants. RSQ responses regarding how the listeners felt about the storyteller’s speech and how similar or different the speech was from their own is shown in table 13.

Table 13. Attitudes toward speech in Malhoriya story

DK DES KAT

N= 10 12 11

How did you like his speech?

Good 70% 83% 64%

Fine 30% 17% 27%

Bad – – 9%

Is the language in this story…from the language

spoken here?

Same 10% 50% –

A little different 90% 42% 45% Very different – 8% 55%

The majority of participants reported the Malhoriya storyteller’s speech was “good.” Attitudes were least positive among Kathariya respondents. Despite this, the majority of Kathariya respondents said the speech was “good.”

Similar to what was said of other varieties, the majority of respondents identified that their speech is, to some degree, different than the storyteller’s. All Kathariya and 90% of Deukhuri respondents identified that the speech was “different.” Again, similar to other varieties, despite these identified differences, positive responses were given regarding the speech sample. This suggests positive attitudes because, though differences exist between the speech varieties, they still view the variety favorably.

Most (69%) Malhoriya respondents said the Deukhuri story was the “same” as their language (table 9) while only 10% of Deukhuri respondents said the same about the Malhoriya story (table 13). This data suggests a tentative analysis that Malhoriya speakers might understand Deukhuri more than Deukhuri speakers understand Malhoriya. But intelligibility testing would be required to conclusively evaluate that possibility.

When asked, most respondents who reported Malhoriya was different in some way said it was because of pronunciation.

The majority of Desauriya (92%) and Deukhuri (80%) respondents said they would allow their child to marry someone that speaks like the Malhoriya storyteller. Sixty-four percent of Kathariya respondents said the same.

After listening to the Malhoriya story, Deukhuri, Desauriya and Kathariya respondents were asked, “How much of the story did you understand?” Table 14 displays their responses.

Table 14. Perceived comprehension of Malhoriya story

DK DES KAT

N= 10 12 11

How much of the Malhoriya story did you

understand?

All 80% 100% 91%

Most 10% – 9%

Half 10% – –

Less than half – – –

As table 14 shows, the majority of respondents reported understanding all of the story. Consistent with RSQ data, Kathariya participants reported through dialect mapping that, after their own variety, they understand Malhoriya best.

During informal interviews, four (of 12) Kathariya respondents reported that Malhoriya is the least pure Tharu variety.

In summary, the majority of respondents for both the RSQ and informal interviews reported they felt that Malhoriya speech is “good.” Participants of the RSQ were able to identify differences between their own speech and the Malhoriya story most of the time. The majority of participants said they would allow intermarriage with Malhoriya speakers, which suggests a lack of strongly negative attitudes. Kathariya subjects appear to view themselves separate and distinct from Malhoriya.

5.4 Summary of attitudes and emic perspectives

In summary of attitudes by varieties toward Dangaura, Deukhuri and Malhoriya, it can be

generalized that no negative attitudes are held toward Dangaura or Deukhuri and mixed attitudes were reported toward Malhoriya.

More RSQ respondents said they understood “all” the Deukhuri story than the Dangaura or Malhoriya story. Subjects from all varieties feel they understand one another to some extent, with the greatest differences being between Dangaura and Kathariya. Kathariya may be more similar to Malhoriya than to Dangaura or Deukhuri. Malhoriya respondents identify Deukhuri as more similar to their own speech than Dangaura.

Differences were reported between speech varieties to varying degrees, but despite the identified differences, overall attitudes appear favorable.

Kathariya respondents often recognized differences between their speech variety and the other varieties of the survey. They often stated that their own speech is “very different” from that of the story’s they heard from other varieties. Kathariya language assistants responded more favorably to Deukhuri speech than they did to recordings of Dangaura and Malhoriya varieties.

6 Language vitality

One way of measuring language vitality is through assessing the function of the language and the degree of interruption of intergenerational transfer. The Expanded Graded Intergenerational Disruption Scale (EGIDS) is often used to classify the level of vitality held in a community or language using a 0–10 scale with 0 representing strong vitality and 10 being extinct (Lewis and Simons, 2010). Through investigation of various factors effecting language vitality, Dangaura Tharu (and its sub-varieties) is assigned an EGIDS level of 5: Written; The language is used orally by all generations and is effectively used in written form in parts of the community. Kathariya Tharu is a level 6a: Vigorous; The language is used orally by all generations and is being learned by children as their first language. A complete description of EGIDS is in appendix G.

Several of the indicators of language vitality that this survey examines are based on Lynn

Landweer’s Indicators of Ethnolinguistic Vitality (IEV), in which she describes several factors that have been “useful in indicating the probable direction a speech community will go relative to the maintenance of, or shift from, its traditional language” (Landweer 2000). This section discusses seven indicators of vitality: domains of language use, intergenerational transfer, group dynamics, network of social relations, language prestige, external recognition and economic base.

6.1 Domains of language use

One aspect of vitality according to the IEV scale is which languages are used in various domains of life. Language use in the home is viewed as one of the primary domains for vitality because the language of the home is generally a child’s first language. Ninety-eight percent of respondents reported that parents in their village usually use the mother tongue with their own children. All respondents with children reported that all of their children speak their mother tongue. This suggests high vitality for the mother tongue in these communities.

Using informal interviews, language choice in 12 domains was investigated. Responses, stratified by age and gender, are displayed in table 15.

Table 15. Domains of Tharu use stratified by gender and age

Men Women

Young Old Young Old Average

N= 13 12 14 11 50

family gatherings 100% 100% 100% 100% 100%

village meetings 91% 92% 86% 100% 92%

quarrelling 62% 100% 100% 100% 91%

scolding/taboo words 62% 100% 93% 100% 89%

storytelling 69% 83% 79% 100% 83%

joking 69% 83% 77% 100% 82%

discussing/debate 62% 83% 79% 100% 81%

telling stories to children 69% 83% 64% 100% 79%

puja/prayer 46% 73% 64% 82% 67%

bargaining/shopping 23% 67% 50% 100% 60%

singing 18% 73% 29% 100% 55%

singing at home 23% 58% 38% 91% 53%

Tharu was the primary language respondents claimed to use in these 12 domains, followed by Nepali. Clear distinctions in patterns of language use are seen when viewing the domains by age and gender. Older women (age 35 and above) report the highest use of Tharu in every domain. Older men also report high use of Tharu in most domains. Young men (age 15–34) report lower use of Tharu than the other demographic groupings in all but three domains (telling stories to children, family gatherings and village meetings). One possible explanation for lower use of Tharu by young men is their high level of education. Only one young man from our sample is uneducated.

The three domains with lowest Tharu use were: bargaining/shopping (60%), singing (55%) and singing at home (53%). This is likely due to high exposure to non-Tharu music (mainly Nepali and Hindi) as well as frequent contact with non-Tharu speakers in bazaars. Although these domains had the

lowest percentages of Tharu use by participants, the percentages still show strong Tharu use given the predominance of Nepali and Hindi in these domains.

One of our goals is to consider language choice patterns in each Tharu variety of this survey. Table 16 shows responses for Tharu language use in twelve domains, stratified by variety.

Table 16. Domains of Tharu use stratified by variety3

DK MAL DES KAT

N= 12 14 12 12

family gatherings 100% 100% 100% 100%

village meetings 92% 100% 75% 100%

quarrelling 100% 100% 73% 83%

scolding/taboo words 100% 100% 67% 83%

storytelling 75% 100% 67% 83%

joking 100% 100% 42% 83%

discussing/debate 92% 93% 50% 83%

telling stories to children 83% 86% 50% 92%

puja/prayer 91% 71% 25% 75%

bargaining/shopping 67% 57% 50% 58%

singing 58% 50% 36% 67%

singing at home 64% 50% 25% 67%

Respondents in the Deukhuri, Malhoriya and Kathariya villages all use their mother tongue relatively often in each of the twelve domains. Desauriya respondents had the lowest reported use of Tharu, when compared to those from other varieties. This was true of every domain except for that of family gatherings, in which all respondents from each area primarily use Tharu.

One of the strongest influences of non-mother-tongue use is in the domain of education. Eighty-five percent of respondents report that children use Nepali in school and not Tharu. This percentage may be even higher because some respondents reported Tharu use in schools by children with children instead of reporting what language was used for instruction (the question was unclear). Only two of 36 parents reported that their children learned Nepali at home. Despite high use of Nepali in school instruction, many respondents report that their children use Tharu outside the classroom at school (i.e. with other children on the playground and by teachers for clarification of instructions). Eighty-nine percent of parent respondents report that they usually use their mother tongue to discuss education matters with their children.

6.2 Intergenerational transfer

An important aspect of language vitality is the extent to which the mother tongue is being passed to the next generation, known as intergenerational transfer.

The Tharu communities of this survey appear to have high intergenerational transfer from parents to children. Over 80% of interview respondents with children report that they use their mother tongue with children in all areas of life. School instruction is the only area they report their children consistently use Nepali. Some examples of this pattern:

• 86% of respondents with children report usually using their mother tongue to tell stories to children

• 88% report that their children usually use their mother tongue when speaking with neighbors

• 86% report that children usually use their mother tongue when playing with other children

• As mentioned in section 1.4.2, 98% of respondents (not just parents, but all respondents) report that their mother tongue is the language most commonly used by parents in the village with their children.

Observations in each location confirm the reported use of Tharu by children.

Consistency of high Tharu use is found when examining intergenerational transfer by variety. Informal interview responses to the question “Do young people in your village speak your mother tongue well, the way it ought to be spoken” are shown stratified by variety in table 17.

Table17. Young people “speak well” responses by variety

DK MAL DES KAT

N= 12 14 12 12

Respondents reported young people as speaking their mother

tongue well.

75% 100% 92% 92%

Table 17 shows that the majority of respondents from each variety report that children are speaking the language the way they feel it ought to be spoken. In addition to current mother-tongue use, the informal interview schedule inquired about attitudes toward what the respondents think will happen with their language in the future. All Malhoriya and Kathariya respondents said they believe their children’s children will speak their mother tongue. Similarly, 92% of Desauriya and 83% of Deukhuri respondents reported the same. This reveals a strong belief in the continuation of their mother tongue.

Based on informal interviews and observations, we found that most mothers speak limited or no Nepali. However, as education increases for women (71% of young women respondents are educated, but none of the older women respondents), so does their ability in the language of education, namely Nepali. Forty-five percent of older women respondents said they speak at least some Nepali. By contrast, all young women reported that they speak Nepali. As this educated generation ages and has children and grandchildren, further studies should be done to re-investigate how this affects the intergenerational transfer of Tharu.

Intergenerational transfer is a critical part of language vitality and, according to the responses from the interviews, the Tharu varieties covered in this survey are being passed on to the next generation. With increasing levels of education, currently all in Nepali, this indicator will require monitoring as young people begin to make choices about which language they will speak with their children.

6.3 Group dynamics

Group dynamics is another indicator of language vitality. “One of the ways that the core of fluent speakers is either supported or undermined is through the language use characteristics of those who immigrate to a speech community” (Landweer 2000). Intermarriage between groups was observed during the survey and, in several instances, the spouse who immigrated learns the local language. In Ghumna, a Deukhuri village, we were told that outsiders who marry-in are not forced to learn the language, but usually do because it is natural when you live there. We observed this in Belganar, a Dangaura village, where a non-Tharu woman married into the community and “fluently” speaks the local variety of Tharu. An additional example is from Ghiya, a Malhoriya village, where our non-Tharu driver spoke Tharu with a village man for directions to a village, when they could have both switched into Nepali instead. These observations suggest strong vitality.

6.4 Network of social relations

Another indicator of language vitality involves language use in a community’s social networks, defined by Landweer as “intracommunity interdependence with dense, multiplex networks utilizing the local language to meet communication needs” (Landweer 2000). A dense network is one in which everyone that is linked or connected to an individual is also linked to one another. A multiplex network is one in which a person is connected to other people through multiple connections (for example, their neighbor may also be their child’s teacher as well as the wife of their doctor). Local language use in a community that has a both dense and multiplex social network indicates high vitality.

Observations allowed for little data collection in regards to the network of the Tharu communities. The most telling information came from interviewing leaders in each community. According to

community leaders, most of the villages have high percentages of Tharu population. In Majgau, a Deukhuri village, it was reported that approximately 3,980 out of the 4,000 villagers are Deukhuri. It was reported in Pabera that approximately 60% of their village’s population is Kathariya; the remaining 40% being Chaudhary (Dangaura). However, it was also reported that around 75% of the villagers speak Kathariya because they said some Chaudhary people also speak Kathariya. In Ghiya, it was reported that all 1,300 people of the village speak Malhoriya, even the Pahadis (hill people) that live there. The Belganar leader reported similar information saying that 800–900 of the 1,200 villagers are Dangaura. By contrast, the Chapargaudi leader reported that approximately 10% of their village is Desauriya.

When leaders were asked about schools and the ethnic background of teachers, information greatly varied. In Ghumna (Deukhuri), 10 of 13 teachers are Tharu. Similarly, a high percentage of teachers in Majgau (Deukhuri) are Tharu. Other locations, however, such as Belganar (Dangaura), reported few Tharu primary school teachers and no Tharu secondary teachers. In Ghiya (Malhoriya), primary school teachers are mostly Tharu and secondary teachers are mostly non-Tharu. In several villages, Tharu medical personnel live and work in the village clinic.

The data collected with regards to this indicator is insufficient to make any strong statements of how the social networks of these communities affect language vitality. However, it would appear that each individual location varies greatly from one another.

6.5 Relative prestige

The relative prestige of the mother tongue within a language community influences the vitality of that language. If a language is perceived as prestigious, when given the choice, speakers are more likely to use it. Several factors affect the prestige of a language including the language of education, religious activities, trade, radio and what is available in a language, as well as government recognition and the presence of organizations (internal or external).

High prestige is found in all varieties through responses to the informal interview question “What language should children speak first?” All respondents from Deukhuri, Desauriya and Malhoriya answered that their mother tongue should be children’s first language. The majority of Malhoriya respondents said the same. Only one respondent reported that English should be first.

There are several organizations located within the Tharu speaking area of this survey, including Kathariya Samaj Nepali, a cultural preservation organization located in Pabera, Kailali district; Help Society Nepal, which works with various types of Tharu development in Belganar, Dang district (their office is located in Tulsipur), and Chaudhary Rin-Badhet Samuha, an economic development

organization located in Chapargaudi, Banke district. Our host in Ghumna had worked with several different organizations (USAID, as well as organizations from Holland and Japan) in order to fund schools to be built. In Pabera, Kailali district, a street drama was performed in the Kathariya Tharu language.

There are radio programs in Tharu. Through asking knowledgeable insiders as well as community members, it appears that Tharu radio programs are available in Dangaura, Deukhuri and Kathariya varieties, but not in Desauriya or Malhoriya. All Deukhuri and Kathariya respondents said they listen to the radio programming available in their language. The majority of Desauriya (88%) and Malhoriya (79%) respondents, whose programs are only available in Dangaura speech variety, reported that they

Expanded Graded Intergenerational Disruption Scale (adapted from Fishman 1991)

LEVEL LABEL DESCRIPTION UNESCO

0 International The language is widely used between nations in trade, knowledge

exchange, and international policy. Safe

1 National The language is used in education, work, mass media, and

government at the nationwide level. Safe

2 Regional The language is used in education, work, mass media, and

government within officially recognized regions of a nation. Safe

3 Trade The language is used in work and mass media without official status

to transcend language differences across a region. Safe

4 Educational The language is vigorous and literacy in the language is being

transmitted sustainably through a system of public education. Safe

5 Written The language is vigorous and is effectively used in written form in

parts of the community though literacy is not yet sustainable. Safe

6a Vigorous The language is used orally by all generations and is normally

learned by children as their first language. Safe

6b Threatened

The language is still used orally within all generations but there is a significant threat to sustainability, particularly a break in

transmission to the next generation by a significant portion of the child-bearing generation.

Vulnerable

7 Shifting The child-bearing generation can use the language among themselves but they do not normally transmit it to their children.

Definitely Endangered

8a Moribund The only remaining active speakers of the language are members of the grandparent generation.

Severely Endangered

8b Nearly Extinct The only remaining speakers of the language are elderly and have little opportunity to use the language.

Critically Endangered

9 Dormant There are no proficient speakers, but some symbolic use remains as

a reminder of heritage identity for an ethnic community. Extinct

10 Extinct No one retains a sense of ethnic identity associated with the

0. International

For this level, we are taking the United Nations as the authority. There are six languages that are recognized as official for this body—Arabic, Chinese, English, French, Russian, Spanish—and these are the languages we place in this category. Other languages used across a number of countries (like Portuguese) are classed as a national language in multiple countries.

1. National

The primary component of Level 1 status is that the written language is used to conduct the business of national government. This need not take the form of being declared “official” in law. A language that is used nationally for oral communication, but which is not used in writing to record the laws of the land or the decisions of the courts, is classed as Level 3.

2. Regional

As with Level 1, the key defining characteristics for this level is use in written form to conduct the business of government. At Level 2 the government in focus is not the national government, but the government of an officially recognized administrative subdivision of the country (for instance, a province or state).

3. Trade

A Level 3 language lacks such recognition as a language for conducting the written business of government. It is still “vehicular”, however, and used by native speakers of other languages across a region for purposes of work or mass media. The general pattern in the EGIDS is that each level adds to what is true in the next lower level; this is the one point where an exception is possible. It is not a requirement of Level 3 that it also be used in formal education as in Level 4. The key component here is verhicularity (that is, being used widely by people who speak different first languages). The original logic of the GIDS as defined by Fishman is that successively higher levels are stronger and less susceptible to language shift. Clearly an unwritten trade language with millions of speakers is in a stronger position than a written local language with only thousands of speakers, even if the latter has achieved formal support in public education.

4. Educational

A Level 4 language is one that is vigorous and non-vehicular and that has achieved sustainable literacy. The fact of educational use of the written language is not enough to qualify as Level 4 (as was implied in the definition in the published paper). Rather, all five of the FAMED conditions as spelled out in the Sustainable Use Model should be in evidence:

• Adequate vernacular literature exists in the domains for which vernacular writing is desired. • Vernacular literacy is being taught by trained teachers under the auspices of a sustainable

institution.

• Members of the language community perceive the economic, social, religious, and identificational benefits of reading and writing in the local language.

• Official government policy calls for the cultivation of this language and cultural identity and the government has put this policy into practice by sanctioning an official orthography and using its educational institutions to transmit local language literacy.

• Members of the language community have a set of shared norms as to when to use the local language in writing versus when to use a more dominant language.

If there is a significant lack of any of these conditions such that removing the government support for education in the language would likely lead to the immediate disuse of literacy, then literacy should still be viewed as incipient and the language should be classified as Level 5.

5. Written

A Level 5 language is a vigorous language in which literacy is incipient. The mere fact that somebody has devised a writing system or even produced a piece of printed literature is not enough to lift a language from Level 6a to Level 5. Rather, the definition requires that some segment of the language community is effectively using literacy in the language. If this is true, but oral language use is

significantly threatened, then the language should be classified as Level 6b. That is, if by ignoring the factor of literacy, it is clear that the language would be classified as Level 6b rather than 6a, then the overall assessment should be 6b since the language cannot be considered vigorous.

6a. Vigorous

A Level 6a language is an oral language that is maintaining sustainable oral use among all generations in the home domain. The most salient indicator of this level is the fact that the language is being

transmitted to all children in the home. By “all” children we do not mean literally 100%, but that it is the societal norm and it is typically followed. A few exceptional cases are not seen as a threat to sustainability, but when a significant number of exceptions emerge such that the community becomes aware that the norm is eroding, then there is a threat to sustainability and the language should be classified as Level 6b. While unbroken intergenerational transmission is the primary indicator of Level 6a, it is not sufficient by itself. Rather, all five of the FAMED conditions of the Sustainable Use Model should be in evidence:

• Adequate oral use exists in every domain for which oral use is desired.

• There is full oral transmission of the vernacular language to all children in the home.

• Members of the language community perceive the economic, social, religious, and identificational benefits of using their language orally.

• Official government policy affirms the oral use of the language.

• Members of the language community have a set of shared norms as to when to use the local language orally versus when to use a more dominant language.

If there is a significant lack of any of these five conditions, then sustained oral use is under threat and the language should be classified as Level 6b. For instance, if the community perceives so little value to using their local language that they would immediately begin transmitting the more dominant

language if only they could learn it, then on-going language use is not sustainable and the language should be classified as Level 6b. Or, if the official government policy is hostile toward ethnolinguistic diversity and calls for the elimination or suppression of this language, then on-going language use is not sustainable and the language should be classified as Level 6b.

6b. Threatened

A Level 6b language is one that has started tipping away from sustainability. If an effort can be made to address any of the above conditions that are undermining the local language, then it may be possible to pull the language back toward sustainability; however, in the absence of such efforts, the community will be likely to continue shifting toward greater use of the more dominant language.

7. Shifting

A Level 7 language is teetering on the tipping point of sustainability. It is no longer the norm for parents to transmit the language to their children, but the parents still know the language, so it would be

possible for intergenerational transmission to resume if the community could be convinced of the value of their language. Level 7 does not mean that transmission to children has completely stopped, but that it is now exceptional. Full transmission happens so infrequently that children who are learning the

language will have difficulty finding peers to speak the language with (including a spouse) when they are adults.

8a. Moribund

A Level 8a language is still in everyday use in some homes, but only among those who are beyond child-bearing age. Thus, the normal cycle of intergenerational transmission has been broken. There may be younger adults who know the language at least somewhat, but they do not regularly speak it with their peers and are not fully proficient.

8b. Nearly extinct

A Level 8b language is no longer used in any home. Those who can still speak the language no longer have a spouse who can speak the language and find little opportunity to use it since there are so few other speakers.

9. Dormant

A Level 9 language is no longer the first language of any living individual. However, it is still the recognized heritage language of an ethnic community and it still serves as a marker of ethnic identity. Some vestiges of the language remain and are passed on within the community to strengthen the heritage identity. This could include names of cultural objects, local place names, traditional greetings, formulaic use of the language in traditional rituals, or even the continuing use of ancient scriptures within in a faith community.

10. Extinct

At this level the language is known only through the historical records. There is no living community that still looks to the language as a marker of its heritage identity.

154

References

Adhikari, Kamal. 2006. Indigenous healing practices among Tharus of Amrai Village in Dang. Education and Development: Research Centre for Educational Innovation and Development 22:106–116.

Blair, Frank. 1990. Survey on a shoestring. SIL International Publications in Linguistics 96. Arlington: The Summer Institute of Linguistics Linguistics, Inc.

Boehm, Edward. 1998. A phonological reconstruction of Proto-Tharu. MA thesis. The University of Texas at Arlington.

Boehm, Kelly. 1997. Language use and language maintenance among the Tharu of the Indo-Nepal Tarai. MA thesis. The University of Texas at Arlington.

Guneratne, Arjun. 1998. Modernization, the state, and the construction of a Tharu identity in Nepal. The Journal of Asian Studies 57:749–773.

Guneratne, Arjun. 2002. Many tongues, one people: The making of Tharu identity in Nepal. Ithaca: Cornell University Press.

Hugoniot, Ken. 1996. Intelligibility and language attitudes among the Tharu dialects of the Western Indo-Nepal Tarai. Ms.

Krauskopff, Gisele. 1995. The anthropology of the Tharus: An annotated bibliography. Kailash: A Journal of Himalayan Studies 17(3 and 4):185–213.

Landweer, M. Lynn. 2000. Indicators of ethnolinguistic vitality. Notes on Sociolinguistics 5(1):5–22. Accessed 2 August 2012. www.sil.org/sociolx/ndg-lg-indicators.html.

Lewis, M. Paul, ed. 2009. Ethnologue: Languages of the World. Sixteenth edition. Dallas: SIL International. Lewis, M Paul and Gary F. Simons. 2010. Assessing endangerment: Expanding Fishman’s GIDS. Revue

Roumaine de Linguistique 55(2):103–120.

McDonaugh, Christian. 1989. The mythology of the Tharu: Aspects of cultural identity in Dang, West Nepal. Kailash: A Journal of Himalayan Studies 15:185–213.

Rajaure, Drone P. 1981. Tharus of Dang: The people and the social context. Kailash: A Journal of Himalayan Studies 8:155–185.

Rajaure, Drone P. 1982. Tharus of Dang: Tharu religion. Kailash: A Journal of Himalayan Studies 9:61–96. van Driem, George. 2002. Languages of the Himalayas. New York: Brill.