LANGUAGE CHOICE AND SOCIAL NETWORKS IN HANUABADA
by
Margaret Mase
A thesis submitted in fulfilment of the requirements for
the degree of Master of Education
Faculty of Education , Northern Territory University
August, 1995 This thesis is dedicated to
the people of Banuabada
NOflTtiEnN TEflfHTOrlY UNIVERSITY L!Or.�RY Northern Territory University
Faculty of Education
To the best of my knowledge and belief the work in this thesis is original, except as acknowledged in the text, and the material has not been submitted in whole or in part for a degree at this or any other university.
Signed:
... Margaret Mase
Date: 24 August 1995
i TABLE OF CONTENTS
Chapter Page 1 Introduction
1.1 The study: a brief description ...••...•••..••.. 1 1.2 Location of the study ...... 2
1.3 Background : languages in contact ••••...... ••. 11
1.4 Statement of the problem and questions arising • 16 2 Review of literature
2.1 Introduction to the review ...... •••...•.•. 21 2.2 Language choice and language shift ..••.•...... 22 2.3 Lan9uage maintenance and language ecology ...... 31 2.4 Soc�al networks ...... •...... • 41 2.5 Patterns of variation and network structure .... 46 3 Methodology
3.1 Aims and objectives ...... 49 3. 2 Data collection . . . . • • . . . . • ...... • . . . . 49
3. 3 Data analysis ...... 63
3.4 Ethical concerns ...... •.•...... •...... • 66
3.5 Limitations of the study ..•...... •..•...... 67 4 Main research findings , Part 1 4.1 Importance of Motu and other languages •.•.••... 70 4.2 Languages spoken at home ...... •..•... 71 4.3 Languages used at school ...... •..... 75 4.4 Languages used at work ...... 78
4.5 Languages used in the city ..•...... •.••.••• 79
4.6 Languages spoken in the church ..•..••...... •... 80
4.7 Literacy ...... 81 5 Main research findings, Part 2
5.1 Language choices ...... 82 5.2 Language choices & respondent characteristics .. 88 5.3 Network villageness & respondent characteristics 99 5.4 Language choice and villageness ...... 104 5.5 Language rank types ...... ••...... •.••.•.• 108 6 Discussion of Results 6.1 Languages at home .....•...... •..•••.•.... 110
6.2 Languages at school ...... •.•.•.•....•...•.... 113
6.3 Languages at work ...... •.•...... •.... 115
6.4 Languages in church ...•...•...... •...... 116
6.5 The language ecology of Hanuabada ...... •...••.• 118
6.6 Speakers' language choices with interlocutors .. 124 6.7 Language choice & res�ondent characteristics ... 127 6.8 Language choice and vlllageness ...... •...... 128
6.9 Language rank types & life modes ••.•...••..•... 129
7 Conclusion ...... 135 Bibliography ...... 139
Appendices ...... 151
ii LIST OF TABLES AND FIGURES
Tables Page
3.1 1990 census figures for Hanuabada .••..•••..•..••. 55 3.2 1990 census figures compared with 1992 count ..•.. 56 3.3 Sample 1 and 1990 census population figures ..•. .• 59 3.4 Sample 1 and 1990 census age group figures .....•. 60 3.5 Sample 2 and 1990 census population figures ...... 62 4.1 Reasons for learning Motu at horne ...... •....•.. 74 4.2 Reasons for learning English at horne ..•..••...... 74 4.3 Attitudes to Tok Ples education ..•...... 77
4.4 Bow literacy was aquired ...... •...•....•.. 81 5.1 Scalogram of language choices .....••...... •.. 84
5.2 Language ranks by respondents ...... •..•...... 86 5.3 Scalogram showing language types ...... •..... 87
5.4 Characteristics of respondents (ranked) ...... 89 5.5 Language ranks by age of respondents ....•...... • 90
5.6 Language ranks by gender of respondents ...... •... 91
5.7 Language ranks by religion of respondents ...••.•. 92
5.8 Language ranks by horne of respondents .....•...... 93 5.9 Language ranks by work of respondents ••...... •.• 94
5.10 Language ranks by home & work of respondents ...•• 95 5.11 Language ranks by early life of respondents ..•... 96
5.12 Language ranks by free time of respondents ...••.. 97 5.13 Language ranks by grade of respondents ..•. .•. . .•• 98
5.14 Network villageness by horne of respondents .•...•. 100
5.15 Network villageness by work of respondents •.•..•. 101 5.16 Network villageness by horne & work of respondents 102 5.17 Network villageness by grade of respondents .•..•. 103
5.18 Comparison of rankings and types ...... •.•..• 105
5.19 Language ranks by villageness of respondents •.... 106
5.20 Language ranks by villageness of network ...... •.• 107
5.21 Characteristics of language rank types ...... 109
Figures
3.1 Age groups to which respondents belong .•...... 61
4.1 Languages spoken at horne .•...... •...... 71
4.2 Reasons for Motu being used at horne .•••.•.•.•.... 72 4. 3 Reasons for English being used at horne ...... 73 4.4 School grades achieved by respondents ...... • 75
4.5 Languages used at school ...... •••• 76 4.6 Combinations of languages used at work ...... •.•. 79
6.1 Languages used or to be learned at horne .....•.•.. 112
6.2 Languages used or to be learned at school .•...... 114 6.3 Languages used at horne, school , work and city ...• 121
iii LIST OF MAPS AND PHOTOGRAPHS
Maps Page
1.1 Papua New Guinea ...... •...... 3 1.2 Port Moresb¥ and environs .•...... •.•...... 4 1.3 Motu and Ko�ta villages •.•...... •••••.....•••••• 5 1.4 Sketch of Hanuabada village ...... •...... •..•. 6
Photographs
1.1 Part of Banuabada village ...... • • • • ...... • . 7 1.2 A Hohodae pier in Hanuabada village ...... •...... 8 1.3 Banuabada village ...•.•..•.•...... ••.•.•..•.. 9 1.4 Betel-nut sellers in Banuabada village ...... 10
TABLE OF APPENDICES
Appendix Page
1 Questionnaire on Lan9ua9e Use ...... •..• 151 2 Respondent Character�st�cs (Part 1) ...•...... •... 155 3 Languages Used in Main Situations ...... 157 4 Languages Learned and Importance .....•....•...••.• 159
5 Reasons for Languages Used ...... •••. 160 6 Reasons for Importance & Learning ...... •...... • 161 7 Questionnaire on Lan9ua9e Use & Networks ....•.••.. 162 8 Respondent Character1st�cs (Part 2) ••••••••••••••• 165
9 Languages Used With Various Interlocutors .•.•••... 166
10 Villageness ...... 168
iv ABSTRACT
This thesis reports the results of a study on the language choices, language attitudes and social networks of the inhabitants of Hanuabada , a multilingual community in Papua New Guinea .
The responses which compr1se the data of this study are the self-reports given by the people of Hanuabada . Data were elicited by means of two questionnaires. In many cases, the questionnaires formed the basis and written records of the interviews conducted with the respondents. Some respondents preferred to complete the questionnaires without interview.
The first part of the study was designed specifically to investigate the language use and attitudes of people in Banuabada. The sample for the first part of the study was selected using stratified systematic sampling, with the clans as strata. There were 55 people in sample 1: 30 males and 25 females.
The second part was designed specifically to investigate the villageness of the respondents and their social networks , 1.e., the number of people they came in contact with most often and how the respondents related to each of them. The sample for the second part of the study was chosen to ensure
v that respondents had social networks across the whole of the villageness and language-use ranges. There were 52 people in sample 2: 29 males and 23 females.
The results show that there is little support in Hanuabada for the Papua New Guinea Education policy on the introduction of Tok Ples (village language ) programs into schools; most of the respondents in Hanuabada favour learning English, whereas only a few favour learning the indigenous language (Motu), at school . A similar attitude seems to exist with regard to which languages should be learned at home . Although 97 per cent of the respondents reported 'using' Motu at home, only 67 per cent thought it should be 'learned ' at home, and whereas only 40 per cent 'use' English at home , 72 per cent indicated it should be 'learned ' at home . The reasons for learning and using English given by the respondents 1n this study can all be classed as 'instrumental' and be seen to relate to social mobility and improvement in employment prospects in the formal sector.
However, Motu 1s viewed as of 'great importance' by most
(83 per cent ) of the respondents . Over two thirds of the respondents in the sample gave reasons for using Motu that implied a sense of both personal and cultural identity - "this is our own language".
V1 After ranking respondents according to their language choices, three broad 'types ' of language users were found. Associated with each language type is a 'life-mode ', which determines the social, cultural and linguistic behaviour, and also the networks of the individuals of that mode . 'Life-mode L' comprises those who live (and most probably work) out of the village, and who belong to the low-rank language type (who use Motu on its own the least ). 'Life-mode M' comprises those who live in the village, but work out of the village, and who belong to the mid-rank language type . 'Life mode-H ' comprises those who live in the village and do not work, and who belong to the high-rank language type (who use Motu on its own more than the other types, and do not use any other language alone) .
vii ACKNOWLEDGEMENT
Without the help of so many people this study would not have been possible . Firstly, to all those people who willingly gave their time with interest and enthusiasm to participate in this study, I am deeply grateful .
For his on-going challenging encouragement and support I thank my supervisor, Dr Brian Devlin. I am grateful to the Education Faculty, the Post Graduate Association and the Post Graduate Office of the Nortern Territory University, particularly for assistance with funding to travel to Papua New Guinea for data-collection and to travel to the Seventh International Conference on Austronesian Linguistics (7- ICAL) in Leiden, Holland.
I thank Dr Bob Litteral (of the Language Curriculum Unit of the Papua New Guinea Department of Education) for his interest and support which enabled me to incorporate the Viles Tok Ples Skul language policy in my data. In the Papua New Guinea National Statistics Office , I am grateful to Mr Boe Douna for his help with the 1990 census figures .
I express appreciation to my family and friends in Papua New Guinea who tirelessly helped me throughout, especially with the painstaking job of collecting the necessary data. Last but not least, I thank my husband and my son, both to whom I owe my inspiration .
viii 1
Chapter 1
INTRODUCTION
1.1 The study: a brief description
This study investigates language use, language attitudes , language choices and social networks of the people of the village of Hanuabada, a multilingual community in Papua New Guinea .
The study is divided into two main parts. Part one examines attitudes and views on Motu usage in the village and reported language use in certain places of one's daily activities , e.g., workplace, school, and home . Part two examlnes social networks and their relationship to language choice. A simple villageness-urbanness scale is used to measure correlations with language choice or language rank.
The responses which comprise the data of this study are the self-reports given by the people of Hanuabada . Data were elicited by means of two questionnaires which were administered in the village during 1992. In many cases, the questionnaires formed the basis and written records of the interviews conducted with the respondents. Some respondents preferred to complete the questionnaires without interview. 2
1.2 Location of the study
Papua New Guinea, with a land mass of 462,840 square kilometres , is the world 's second largest island. It lies north of Australia and west of the Solomon Islands , between longtitudes 141° and 160°, and latitudes 1° and 12° south. (See Map 1.1. ) According to the 1980 National population census (National Statistics Office , 1985 ), 82 per cent of the population over 10 years of age lived in rural villages, 12 per cent lived in urban areas and the remainder were in non-village settlements such as plantations and mining camps. The 1990 National population census (National Statistics Office, personal communication, 1994 ) reveals little change in these statistics.
Hanuabada - the location of this study and the place where I
was born and raised - lS a large village on the south coast of Papua New Guinea (see Maps 1.2 and 1.3). It is often described as an "urban village" because Port Moresby (the capital city of Papua New Guinea) has developed around the village . Most of the inhabitants of Hanuabada still live in houses built on stilts over the sea with access provided by means of piers (see Map 1.4 and Photographs 1.1 to 1.4). 3
1 ·.
\ \ . .\ . I � \I --:1� . � .. '" ' �t
o· .. '
� ......
., J � - ! • - r_ ' ' - � 3 ,. \: �r) ·.r. �< �..) ::::.
� Ii I 4 "
.. C"l . 5
0 10 20
KILOMETRES
Korobosea')_� Paf'\ • • · I V. · · .' Tubusere•a
·.:-:-.-.-:-:·� ·. · ·. Bankau
• r-"�tu Vollages • 1(�,� Vol'agcs
Gabagaba
Hap 1.3: Motu and Koi ta vi llages Banuabada is an urban village situated on the Port Moresby harbour foreshore . o 0o 0 0 0
0 o 0 o o 0 � 0 0 0 ""' toe�' O o o o 0 O oO 00 o o o 0 o o o o 0 � \o ci\Y 0 0 0 0 0
··. - .. · · r ... . ·. · ·· . · · .·· ·· .. · . · ·· ·· .. ea · s · · . ·· .. ·. ..· · · .. : .. · . ·· . ·· ...... a ...... Mavat _ a e · \ubumag j hoda hoi _. \-\o au Va Gunina i Ap dubuna Sota Kwat� amona Kahan 0'1
,,.y//erge nuerberder \ Her Sketchot MerP 1. 4: 7
Photograph 1.1: Part of Hanuabada village
In the background can be seen new, multi-story residential units. 8
Photograph 1.2: A Hohodae pier in Hanuabada village
This photograph shows a p1er belonging to the Hohodae clan. The pier is used to gain access to the houses built over the sea. 9 .
Photograph 1.3: Hanuabada village
Toua Guba Bill, a prestigious residential suburb for the townsfolk, can be seen in the background. 10
- � �-:-- ' - I
Photograph 1.4: Betel-nut sellers in Hanuabada village
The people in the picture are locals who buy betel-nut in bulk from the main markets in the city and sell it dearly to the villagers. The market for betel-nut is always favourable in the village. In some cases this is the only way people can earn money to survive . 11
1.3 Background: languages in contact
The Asia-Pacific region is linguistically diverse . This can be seen clearly by considering the following figures for languages per country , noted by Dixon (1991). In the Philipines there are 60 million people speaking a total of 160 different languages. In pre-European Aboriginal Australia it is estimated that approximately 250 languages were spoken by between one and two million people. In Indonesia, excluding Irian Jaya, there are 350 languages spoken by a population of 180 million. In Irian Jaya alone 200 languages are estimated to be spoken by 1.6 million people . Throughout the Pacific islands, the population �s estimated to be approximately 2 million and it is believed that there are about 260 different languages in use .
In Papua New Guinea (PNG) there are an estimated 869
languages (Dutton and Muhlhausler, 1991, p . 1) shared by an estimated population of just over 3.9 million. The National Census figures for 1990 show a total population of 3,607 ,954 with an annual growth rate of 2.3 per cent, but this excludes North Solomons Province where the census could not be conducted due to the prevailing political situation at that time (National Statistics Office , 1990). This makes the linguistic diversity in Papua New Guinea the greatest �n
the region, whilst its average of approximately 4500 speakers per language is one of the smaller in the region (see Crowley and Lynch, 1983, for a breakdown of 12
Austronesian and non-Austronesian language types and the number of speakers for each one). It is indeed a 'sociolinguistic laboratory ' (Wurm, 1979).
One of the many languages in Papua New Guinea is Motu , a minor language of the Melanesian sub-group of languages which belong to the Malayo-Polynesian group, commonly known as the Austronesian family of languages (Dutton, 1976a).
The language is spoken by over 23,000 people (Crawford & Lahui, 1980 ) of Motuan and Motu-Koita origin, who are the native inhabitants and traditional landowners of Port
Moresby and surrounding areas. The name 'Motu-Koita ' derives from the broad cultural or language groups Motu and Koitabu which eventually became one sub-group . Approximately 4000 of the Motuan speakers live �n Hanuabada (National Statistics Office , 1990).
Believed to be the largest village cluster in PNG, it is no surprise that Banuabada 1s the largest Motu-Koita village.
Historically, Banuabada consisted of five sub-villages , now referred to as clans, of which two were inhabited by people
of Koita origin and three by Motuans . Subsequently they merged and became one big village with a total of nine clans : Gunina, Hohodae , Apau, Vahoi,Botai , Tubumaga, Kwaradubuna , Havara and Kahanamona (Belshaw, 1957 ). What is now the Bohodae clan (which is the clan to which I belong) used to be one of the Koita sub-villages. 13
Immediately adjacent to Hanuabada is the village of Elevala. Hanuabada combined with Elevala is known collectively as Poreporena . Each section (i.e. Hanuabada and Elevala) has its own church building, but both belong to the United Church. Hanuabada is the section of the Poreporena village that stretches from the Hohodae clan across to the Botai clan (see Map 1.4). Although some people still continue to build their homes over the water, more and more people are now moving onto the land and building there. In the land area of Hanuabada where these houses have been built , I found that several heads of households are actually from Elevala.
The Motu were seaman, traders , potters and fishermen and the Koita were agriculturalists (Seligman, 1904). It is believed that the Motuans inhabited the coastline and the Koita, the hinterland. When they came in contact both groups agreed to be partners in exchange of food. The Motuans would exchange fish and seafood for vegetables like yams and taro as well as game like wallabies and bandicoots . Ultimately both groups intermarried.
Between 10 and 15 per cent of the population in Hanuabada
are Koitabu language speakers but they prefer to use Motu in the village . Koitabu, which is not an Austronesian
language, is strongly alive in other predominantly Koita villages outside Hanuabada; namely, Baruni, Kilakila,
Korobosea, Kouderika, Roku , Papa, Kido, and Gorohu . Baruni became a village by the coast when a large group of Koita
people migrated there . 14
Port Moresby, the national capital of the independent state of Papua New Guinea, with a population of 195,570 and an annual growth rate of 4.69 per cent (National Statistics Office, 1990 ), is made up of local and migrant people from many different cultures, mostly from within Papua New Guinea but also from overseas (7,481 non-citizens ). Since British New Guinea was declared a British protectorate in 1884, and Australia assumed administrative resposibility for it in 1906 , renaming it "Papua", European contact has brought about many sociocultural changes.
Some of the obvious changes, especially for the local
Motu-Koita people, include the transition from a subsistence to a largely cash economy . In the past, people worked gardens to produce their own vegetables and fruits, but today they buy food from the markets and shops.
The sudden migration of people from predominantly rural village settings to what is regarded as the big city is another major change. As Port Moresby is a modern capital city, it is where the government 's central offices, commercial enterprises and most modern types of buildings
and institutions can be found. It is also one of the main
ports in the country; the other main ports are Lae and
Rabaul, located on the north coast of the main island and the easternmost tip of New Britain respectively . The fact that Port Moresby has these facilities is one reason why it has attracted a steady stream of migrants from rural areas. 15
Rural-urban migration and intermarriage are two examples of factors contributing to changes in language use and language choice in Port Moresby . Because of the need to communicate with speakers of other languages , people are forced to use several languages, both in isolation and in combination, thus exercising language choice in the city. For example , a Motuan bank clerk attending to an Indian customer will probably not speak Motu or an Indian language to him or her, but will use a lingua franca which both parties speak and understand. The three lingua francas used in Port Moresby are English, Tok Pisin and Hiri Motu.
English is the official written language and is used in schools, universities, government and business. It is seen as the language for international communication. English is the language of instruction in PNG, therefore it is the language used in the two pre-schools, the two primary schools and the two high schools in and around the village of Banuabada .
Tok Pisin or Melanesian Pidgin is believed to have developed as a language before significant European contact with Papua New Guinea. There is evidence that it was brought to Papua
New Guinea in a stabilised form from Samoa (Muhlhausler, 1979 ). Much of its vocabulary is derived from English
(Muhlhausler, 1979 ). It appears to have a wider use than Biri Motu. As a lingua franca, it seems to be gaining popularity very quickly . People seem to accept the language very easily and adopt it. This is apparent in the speech of
those Motuans from Hanuabada (and from other Motuan villages 16 as well) who go to work or school in the city . Since the language is now being acquired as a first language by some children �n the urban centres and in a few villages, Tok Pisin is becoming a creole (Smith, 1992). But for most people in Papua New Guinea (and this applies in Hanuabada) Tok Pisin is not a first language (Smith, 1995).
From my personal observations pr�or to undertaking this study, it seemed that Biri Motu is not growing as fast as Tok Pisin is, in the sense that many people use the latter as a lingua franca. However Hiri Motu is known by a lot of people from the Papua region. It is an indigenous pidgin based on Motu , which was used between the Motuans and people from the neighbouring language groups well before the missionaries and the colonisers. When Australia assumed responsibility for Papua, the language was adopted as a lingua franca by the then armed native constabulary of Papua. Biri Motu (sometimes referred to as Police Motu ) should not be confused with Motu (i.e . Motu proper ). Although both share 90 per cent vocabulary , mutual intelligibility between the two is predominantly low to very
low (Wurm, 1964).
1.4 Statement of the problem and questions arising
The speech community of Hanuabada , like most other places in Papua New Guinea, �s multilingual. Besides the three lingua francas, English, Tok-Pisin, and Hiri Motu- there are the
local vernaculars, Motu and Koitabu, and the vernacular 17
languages of settlers in Hanuabada, e.g. , Goilala and Kerema. 'Code-switching ', defined as when one or both speakers change from one language or language variety to another in the course of a conversation (Scotton, 1982; Scotton and Ury , 1977 ), is pretty much the norm in the village. It is not considered strange for a speaker to speak Motu, English, and Tok Pisin or any other language for that matter, in the same conversation. Code-switching is more common among the younger people, especially the school-aged.
Personally I know two young families who use Motu predominantly with the extended family and English or Tok Pisin predominantly in their nuclear families. Nevertheless code-switching among these languages
predominate in situations when the two types of families interact, which is fairly often. Although the conversation may involve extensive code-switching, all the participants involved seem to understand each other and communicate effectively.
Due to language contact situations like traditional trade, the arrival of missionaries, and eventually colonisation, Banuabada has undergone rapid social and cultural changes . These changes will have in some way affected the use of the Motu language within the village. Like all other living
languages Motu continues to change and this is evident in
its structure , sound system, meanings and lexicon. For example, about twenty or more years ago people would have
used gaim to mean uncle. Nowadays it is vava. It is also 18 noticeable that h is pronounced by only a few of the older people whereas it is pronounced by many of the younger people in the village. These differences result in distinct language varieties. Many older people (i.e., the seventy plus age-group ) agree that the kind of Motu spoken by children in the village is not exactly the same as their grandparents ' variety .
Before commencing this research project, something I noticed was the use of Motu words by the older people that younger people didn 't understand . I also noticed that grandparents' conversations with other grandparents were predominantly in Motu. I could hear few English or Tok Pisin words in their conversations . However when grandparents spoke to their grandchildren, two or more languages could be heard in the conversations . Children used more words borrowed from English in their speech . These words have been adapted to
fit the sound system of Motu. Each word that is borrowed sounds different and is spelled slightly differently from the original, e.g. sikuli "school'' . These distinct language varieties are examples of age-group-related differences in
language use within the village .
English plays a significant role in a Hanuabadan child 's life because it is the only language that is required to be
learned at school . Children spend a lot of their time
learning and speaking English at school. But English is not
only used at school. Many Hanuabadan 's use English elsewhere, e.g., at work or in the city. Is the village of 19
Hanuabada experiencing 'language shift •, i.e. , the change at the community level from speaking one language to speaking another language?
The consequence of a 'shift' in some bilingual/multilingual situations is that an indigenous language is gradually replaced by an outside language. Is the rapidly changing multilingual community in Port Moresby a threat to the existence and maintenance of Motu in Hanuabada? As the urban city of Port Moresby develops around Hanuabada, the need for acquiring lingua francas may increase the degree of language shift. In many countries where European contact and/or migration has created mainstream societies, language death has occurred (Fesl, 1987 ). Although contact which results in mainstreaming is not the only cause of language death, it is clearly important to consider this factor when looking at language shift in Australia and Papua New Guinea during the last two hundred years . In Australia, for example, many Aboriginal languages have died out and more continue to do so (Schmidt , 1990) . This is the main reason why there is increasing interest in maintaining and
promoting Australia's remaining Aboriginal languages, and �n reviving the dead and dying languages (Devlin, 1986a; HcConvell, 1986; Schmidt , 1985, 1990; Evans , 1985; Amery, 1985).
Language maintenance is the broad topic area within which
this study investigates language use, language attitudes and language choices by the people of the urban village of
Banuabada . The study is divided into two main parts. Part 20 one examines attitudes and views on Motu usage in the village and reported language use in certain places of one 's daily activities , e.g. , workplace, school, and horne . More specifically, it seeks to answer the following questions : 1. In which situations and activities is Motu used, and in which is it not used at all? 2. Which languages do people think children should learn in schools? 3. Which languages do people think children should learn at home? 4. What attitudes do people have towards Motu and Motu language use in the village? 5. What functions or roles does Motu assume in the village?
Part two examines social networks and their relationship to language choice. A simple villageness-urbanness scale is
used to measure any correlations with language choice or
language rank. Chapter 2
REVIEW OF LITERATURE
2.1 Introduction to the review
This study investigates the inter-language variations as demonstrated by the language choices of people in Hanuabada and seeks to relate these to various extra-linguistic features , in order to obtain a deeper understanding of the factors affecting language maintenance in the village .
Thus, in the second section of this chapter I review the literature on language variation, choice and shift.
Language shift (the change-over in a community from speaking one language to speaking another ) results from the loss of
the functions of bilingualism, that is, the loss of functional choices between languages. If one language assumes the functions of another language, then this can lead to 'language death '.
In the language ecology of a multilingual society such as Papua New Guinea, language-switching is a common variation. In the third section, the general literature on language
maintenance and some of the specific literature on the language ecology of Papua New Guinea is reviewed . 22
The fourth section considers the literature on social networks and their relation to member identity and life-modes (these being three different economically determined social cateogories, each with different orientations to work, leisure and family).
The final section reviews literature on the relationships
between language choices and social networks . In particular, it considers the use of implicational scaling to rank respondents according to their language choices . It also considers criteria for measuring network structure .
2.2 Language choice and language shift
From linguistic research (Trudgill, 1983) it is known that
the speech of men may be different from that of women . In
some accents of American English, women are said to have
more peripheral vowels (more back, more fronted , higher or lower) compared to men. These differences are so small they are generally not noticed. Jesperson (1922 ) noted that Carib Indian men use a number of words which only adult
males are allowed to use when preparing for battle . It was
felt that if women or uninitiated men used these special
words it would bring bad luck. Jesperson points out that in
this case the language difference is directly related to
taboos . Taboo may perhaps be one of the powerful influences
on the growth of separate gender-related vocabularies in some cultures (Trudgill, 1983). 23
Aikio (1982) and Lindgren (1984) found that in the groups of Finnish speaking Kven immigrants who have settled in northern Norway the women have become more Norwegianised than the men. To some extent, this indicates that women tend to be more sociolinguistically progressive. However, in Gapun (a small village 10 km inland from the north coast of PNG, roughly halfway between the lower Sepik and Ramu rivers ) women tend to speak more Taiap (the local language and first language for most of the adults) than men do. In their informal conversations with one another, and
importantly in their kroses screamed over the village, women tend generally not to code-switch into Tok Pisin as much as men do . (Kulick, 1992b, p.17 ) In so far as Tok Pisin seems to be strongly associated with the male folk of the village, then they could be said to be
the more sociolinguistically progressive.
One of the issues highlighted in these studies is variation
in the linguistic repertoire of speakers. The subtle intra-language variations (e.g., dialects and accents ) and
the inter-language variations (e.g., code-switching) are
identifying features for each speaker. Milroy (1987a)
investigated intra-language variation in Belfast. She found that speakers used different styles and registers primarily as features for group identification; i.e., a particular style of speech was associated with certain group solidarity. Gal (1979 ) noted both intra- and inter-language
variations in the repertoire of speakers in Oberwart . This
study is further discussed in section 2.4 of this thesis. 24
The study of linguistic variation has been strongly influenced by the work of Labov. He and his colleagues (Labov, 1963; 1972a; 1972b; Labov, Cohen, Robins and Lewis, 1968) have tried to understand how language change happens and what causes it by examining how it is 'embedded ' in the social and linguistic context . It now seems clear that variation in language is patterned, that it is related to language change, and that it depends crucially on the interaction with the
social setting. (Fasold, 1990, p.266)
Language contact is a primary factor in language change (both intra-language and inter-language ) in Papua New Guinea (Collier, 1977; Combs , 1977). In assessing its influence, the frequency, duration and degree of contact are all aspects which need to be considered (Mackey , 1968). Frequent (daily ) contact on a large scale is needed to cause large-scale language change (Collier, 1977; Combs , 1977;
Tabouret-Keller, 1968 ). Collier (1977) suggests that it is reasonable to assume that contact restricted to trade and work contexts will provoke less change than contact in sociable situations.
Lithgow (1992) notes the considerable changes which have occurred over a twent y-five-year period within the languages of the Fergusson and Normanby Islands of the Milne Bay
Province of Papua New Guinea as a result of contact with
Oobu (a local language used as a lingua franca in the islands ). He (Lithgow, 1992, p.37) lists the causes of the extensive changes in the Sanalowa Island dialect of Dobu to 25 be: use of Dobu by the church; Dobu-speaking visitors; intermarriage with Dobu speakers ; presence of many Dobu settlers ; and, the fact that the island has a small population (about 300). Lithgow (1992, p.47) suggests that the effect which Dobu is having on the languages of the Fergusson and Normanby Islands is "directly relateable to the use of Dobu as a church lingua franca and subsequently the general lingua franca for this area" .
Changes have also occurred in the use of the language Abu ' Arapesh in the Sandaun Province of Papua New Guinea as a result of contact with other languages and cultures (Nekitel, 1992). The Abu ' people appear to prefer to use
Tok Pisin rather than Abu ' Arapesh. Nekitel (1992, p.51) gives several reasons: inter-ethnic marriages; children attending schools outside the Abu ' language area; language policy which restricts the use of indigenous languages in official domains; movement of Abu ' people to and from towns or hospitals; parental neglect in not encouraging children to learn their mother tongue .
HcConvell (1991a ) discusses three theories of language shift (the change-over in a community from speaking one language to speaking another) - interactionistjvariationist theory, adaptation theory and domain theory - which all make predictions for language shift in different ways.
'Interactionist/variationist theory ' says that people engage in switching codes not only because of the person they talk to, the topic they are discussing and where they are 26 discussing it , but because they want to express a feeling or a viewpoint (McConvell, 199la) . Milroy (1987a} has stressed that studying the vernacular speech of particular individuals in a community is crucial for understanding language change and language maintenance issues because of the way in which the kind of language used by the vernacular speaker expresses feelings of solidarity with others.
'Language adaptation ' is based on the theory that rapid cultural change is responsible for language shift because traditional languages are unable to describe or talk about modern things, issues and concepts (McConvell, 199la) . The emphasis is on topic more than the setting or situation, or the interlocutor type and his or her identification .
'Domain theory', pioneered by Joshua Fishman (1972b), is the view that language types (distinct languages or varieties ) can be related to distinct domains of social interaction
(e.g. , the home or family, school or education, the neighbourhood ). It assumes that language choice can be shown to be related to the social context (which is broadly defined to include the interlocutor and the topic , as well as the place ) in which the speech occurs. According to this view, a bilingual community is unlikely to undergo any major shifts in language use if each language is restricted to specific domains; i.e., if the community exhibits bilingualism with diglossia (Fishman, 1967 ). In bilingual diglossic situations, one speech variety (the standard or high variety ) is usually associated with high status and has high prestige, while the second variety (the non-standard or 27 low variety) is usually associated with low status and has low prestige (Ferguson, 1959 ; Fishman, 1967). (See also Fishman, 1964, 1965, 1968, 1970, 1971, 1972a, 1972c , 1978, 1989 , 1991. ) For example, Greenfield (1972) showed that the Puerto Rican community in New York city had a diglossic situation in which Spanish was the low language and English was the high variety .
Parasher (1980), 1n his study of educated Indian bilinguals, surveyed reported language use in the following seven domains: f amily , neighbourhood, friendship, transactions, education, government and employment . Friendship, family and neighbourhood were the low domains in which the mother tongue was usually chosen. To Parasher's surprise, English turned out to be the choice in friendship and neighbourhood domains. The explanation was simply that most of the Indian bilinguals in his sample did not have the same mother tongue .
Language choice in a specific domain is not necessarily uniform or absolute in a multilingual community. In some circumstances language choice can best be described in terms of probability; i.e, which language is most often used or most likely to be chosen . Domains can then be ordered according to the degree of probability of the choice of a particular language. For example, Laesa (1975 ) investigated the choice of English and Spanish within the three contexts of classroom, family and recreation activities in school . The groups under study were three Spanish-speaking American groups ; namely, Cuban Americans in Miami, Mexican Americans 28 in Texas and Puerto Ricans in New York city . The study found that the use of Spanish was most often reported in the f amily, less often in the recreation activities and least often in the classroom. In situations like this, where two or more languages compete with each other for the same domain (i.e. , where there is bilingualism and no diglossia) domain theory predicts that language shift will occur (Matthews , 1983). Van Bout and Munstermann (1988) take the absence of consensus about language or variety choice in a domain to indicate that the domain is 'unstable '; i.e., it is in a process of transition toward a stable monolingual or standard-language domain.
Van Bout and Hunstermann (1988) suggest that it is necessary to describe the relations between domains in terms of two dimensions: solidarity and instrumentality. They use these two dimensions to explain the fact that in some domains or situations there is disagreement between informants on the appropriateness of the use of a particular language variety. Thus, they say that the ambivalence they found 1n their informants ' responses for the parent-child situation arises because choices about language appropriateness in this situation are determined by motivations of solidarity on the one hand, and by instrumental (educational ) motivations on the other hand . Ryan (1979, p.152) points out that "the contrast between instrumental and integrative motivation appears to be quite similar to that between status and solidarity". 29
When speakers make a shift in language use (e.g. , adopting English, instead of or as well as the indigenous or ethnic variety, as the 'home ' language ), the low variety of the language used in diglossic/triglossic situations is the one most likely to be under threat . However, although nonstandard speech varieties usually have low prestige, they are also associated with values of importance for an ethnic group, especially those related to group solidarity, comradeship, peer group approval, cultural continuity and intimacy within the group. Mackey (1978) and Ryan (1979) consider that these values and language loyalty attitudes contribute to the continued usage of the language . HcConvell (1986) feels that the social function of relating to social groups and maintaining solidarity is perhaps the most important function of language choice in bilingual situations, and may above all other factors , determine whether or not language shift takes place.
HcConvell (1986; 1991a; 1991b) believes that each of the three theories of language shift outlined above (viz., interactionist/variationist theory, adaptation theory and domain theory) is inadequate in some respect. He suggests that a theory of language shift must recognise that there are three functions of language which influence choice (McConvell, 1991 a) :
the basic 'communication function ' is particularly important in interactions between bilinguals in the
speech community and outsiders who know only one of the 30
languages in the community: the bilingual speaker can
choose to speak a language that the outsider knows or one that he or she does not know;
the 'social function ' influences the choice of
language (s) to be used between bilinguals who share the same language repertoire : a language can be chosen to mark the identity of the speaker, listener or both to a particular social group;
the 'cultural function ' influences language choice where the speaker needs to make available a particular
world-view and system of knowledge, which lS better
expressed in the language of that particular culture .
McConvell (199la ) calls the resulting theory: "the functional choice theory of language shift". According to this theory , language shift results from the loss of the functions of bilingualism, that is, the loss of functional choices between languages, and in
the case of the social function, their likely replacement by choices of style within one language .
(McConvell, 199la, p.lSl )
The insights of the interactional, language adaption and domain theories of language shift can be incorporated into this functional choice theory to provide a powerful framework for the analysis of the multilingual situation in
Banuabada. 31
2.3 Language maintenance and the language ecology of PNG
In many multilingual countries, the adoption of a single national language, or of several national languages , leads inevitably to language shift. Concerned to assist the spread of the official language of Sudan, Jernudd (1968, p.l76) examines the "sociolectal obstacles to the expansion and use of Arabic" by the Jebel Marra For in Sudan. He notes that in the home and for tribal activities, eg ., dancing, the language For is used; whereas in the local market, with a friend, and talking to the sheikh, Arabic and For are used; and, in the town markets or when addressing an unknown person, Arabic is used. He suggests the need for an active campaign to increase the use of
Arabic , so as to achieve a bilingual situation in which "For gets its permanent stronghold in local and tribal use, whereas Arabic would be used elsewhere" (Jernudd , 1968, p.l7 7). Clearly, although the Jebel Marra For were experiencing language shift, Jernudd (1968) suggested that their language For would be maintained.
'Language maintenance' is defined as the degree to which an individual or group continues to
use their language, particularly in a bilingual or multilingual area or among immigrant groups . (Richards, Platt and Weber, p.158) 32
According to Fishman (1989), there need not be a conflict between language spread and language maintenance. However, if the new spreading language assumes more and more of the functions of the local language, then this can lead to 'language death ', i.e. the speakers switch to using another language, which has been in competition with their mother-tongue in
that geographical region. (Dixon, 1991, p.231)
Siegel (1992 ) does not mention the impact on language maintenance when he notes the spread in Papua New Guinea of the language Tok Pisin: as the lingua franca of urban centres; in government ; in the media; and, in education .
Smith (1992 ) examines the prospects for the Susuami language , spoken in the Upper Watut Valley outside Bulolo 1n Papua New Guinea. In 1979 the language was spoken by about fifty people. In a 1980 census of Susuami speakers , Smith (1990) noted that: some inhabitants of Manki village used up to six languages on occasions ; there were only seven households in which both partners were Susuami language speakers , and partners who were non-speakers were reluctant to learn it; primary school children were educated in
English, and used Tok Pisin rather than Susuami when playing . In 1990 there were only fifteen people who claimed to speak the language, and Smith (1992, p.l19) draws the conclusion that "the language is doomed to die with the present generation of adult speakers". 33
Taiap is another example of a language 1n Papua New Guinea which is on the verge of dying. It is spoken in the village of Gapun, between the Sepik and Ramu rivers . The Gapun villagers were multilingual, being able to speak the vernacular languages of the surrounding villages. However, the speech patterns of the Gapun villagers, their ideas about the world and the increasing penetration of the state and the market economy into the community are all contributing to the demise of the vernacular.
(Kulick and Stroud, 1992, p.225)
When Kulick conducted his research 1n the village in 1987 , there were only 89 fluent speakers of the language and all children under ten spoke only Tok Pisin. The latter was introduced by young men returning from the plantations, who used the pidgin language they had acquired to show their knowledge of the outside world and to enhance their reputation in the village . The introduction of Tok Pisin by this method has been noted elsewhere in Papua New Guinea (Laycock, 1979; Mulhausler, 1979; Sankoff, 1976; Nekitel, 1992). In Gapun, Tok Pisin was eagerly adopted by the villagers so as to "facilitate access to the secret underpinnings of white power and wealth" (Kulick, 1992b, p.20). Tok Pisin's identification with power was reinforced by its use by priests in the church and by the colonial authorities. But Kulick (1992a) points out that there was not a rigid diglossic relationship between Tok Pisin and
Taiap. 34
Instead, Tok Pisin became absorbed into the villagers ' talk in such a way that there are constant mixes and switches between that language and the village
vernacular. (Kulick, 1992a, p.253)
Schlieben-Lange (1977) suggests that a language will be abandoned by its speakers if it no longer performs a function either as a 'language of power ' or a 'language of solidarity '. Franklin and Stefaniw (1992, p.S) report that the ritual pandanus languages in the Southern Highlands Province of Papua New Guinea were used by the Kewa and
Imbongu people to control "potentially hostile elements or people" when collecting pandanus nuts in heavily wooded areas. Although some villagers say that they no longer use a pandanus language because they are no longer afraid, Franklin and Stefaniw (1992, p.1) suggest that it "reflects the general apathy of many members of the present generation towards the vernacular and their wholesale adoption of Tok Pisin" . On the other hand, Pawley (1992) reports that the ritual pandanus language of the Kalam people of the Upper Kaironk Valley in Madang Province is apparently stable. One factor in its survival may be that it is used playfully in and around the home as well as seriously in the forest .
The role of the church in culture and language change in
Papua New Guinea has already been referred to. Senft (1992 ) notes that the influence of Christian belief on the concepts of magic and religion has resulted in the loss of the
Kilivila variety biga tommwaya (old people's language) or biga baloma (language of the spirits of the dead) . Renck 35
(1992), on the other hand, suggests that the character of religion in many indigenous societies of Papua New Guinea was doomed to disappear anyway . He maintains that where indigenous languages (e.g., Yagaria in the Eastern Highlands Province ) have been used in the introduced and compartmentalised religious domain, they may have changed considerably but at least they have a use and are kept alive.
Smith (1992 ) describes the multilingual village of Musom in which the dominant language is Musom, but people also speak
Tok Pisin, the Nabak and Duwet languages of neighbouring villages, and Yabem (the language used by the Lutheran mission). In spite of the fact that the Austronesian language Musom only has about 230 speakers and that nearly half the households contain marriage partners of other languages, it appears to be relatively healthy. This is attributed to the fact that in the case of these mixed marr1ages only marriage partners are welcome to reside in the
village - their unattached relatives are not permitted to settle there . There is a pride in the language and
identity of the village and an ideology that newcomers
must adapt to village norms of language and culture. (Smith, 1992, p. 119)
However, Smith cautions that the future of Musom is not assured. Increasing inter-ethnic marriage, use of Tok Pisin in church and with non-Musom speakers, visits to the market and hospital in Lae, and the English-only policy in school, may all lead to a decline in the use of Musom. 36
Dressler and Wodakleodolter (1977) suggest that language death is associated with 'secondary socialisation of children in the second language ' - i.e. , when the mother tongue is used only in the home and the second language is used everywhere else. For Matthews (1983), signs characterising a dying language include : smallness of group; polluted accent and lexicography ; lack of language vitality; decreased use by younger generations ; absence of diglossia; and , emergence of semi-speakers . Above all else, "the most crucial factor �s whether or not the parents are teaching their children the language" (Matthews , 1983, p. 13 ).
Devlin (1986b) suggests two other interpretations for the term "language maintenance ", in addition to the sociolinguistic dictionary definition given above :
In the bilingual education domain , maintenance generally implies continued support for the initial language of instruction through the primary years of
schooling. A broader and more political application of maintenance involves any degree of official support the
vernacular language receives. (Devlin, 1986b, p.2)
Although the Administrative Officer of Papua New Guinea in the early twentieth century, W.C. Groves, claimed that it would not be possible to teach English in the village schools and that therefore local languages should be used
(Johnson, 1977), there appears to have been little official support for vernacular languages . The 1955 policy of
"universal literacy in English" can hardly been seen as 37 support (Johnson, 1974). Not only was the policy not supportive of vernacular languages, it was also not very helpful for the education processes required in the classroom. Kale (1990 ) suggests that in spite of the large amount of time allocated to the teaching of English, it is debatable whether the majority of children are being 'educated ', or merely half-learning a language of dubious usefulness to them in the immediate future or
the long run. (Kale, 1990, p194) The Educational Sector Review of 1991 agrees, pointing out that the requirement that all children acquire initial literacy in English has resulted in "primary school leavers remaining functionally illiterate in any language " (Papua
New Guinea Department of Education, 1991, p.4).
Thurston (1992) is concerned that widespread literacy �n English could lead to the abandonment of vernacular languages. Some linguists, e.g. , Ahai (1984), Olsson
(1984), and Wurm (Swatridge , 1985), argue that a first language should be used in the beginning years of education in Papua New Guinea . Smith (1992 , p. 121) supports the introduction of vernacular literacy programs , which could achieve the aims of "improving literacy skills, transmitting traditional cultural values and maintaining language ". Kale (1990) suggests that Tok Pisin should also be used as a medium of instruction in bilingual education programs . Hall
(1955 ) supports the use of Tok Pisin in early schooling, suggesting that it could then be used in the subsequent teaching of English. 38
Litteral (1975) proposes that for the first two or so years of school, education should be in the child 's first language. For the next three years , he proposes that one of the lingua francas , Hiri Motu or Tok Pisin, should be used, with English taught as a separate subject . By the end of primary school, teaching would be in English, and the other languages would then be taught as separate subjects. Dutton (1976b) suggests that the national language (he proposes Tok Pisin) be introduced after the children have acquired literacy in their mother tongue, with English being introduced at the end of secondary school . More recently,
Kerema (1989 ) proposes several language models for schools: 'mixed medium' in which pupils and teachers switch between a number of languages ; 'dual medium ' in which a language would be assigned to each subject; and , 'parallel medium ' in which pupils are streamed according to the language they know best , with another language being taught as a subject . Having a variety of models to choose from recognises that vastly different linguistic conditions exist in the various parts of Papua New Guinea and that no single model will work everywhere .
In 1989, the Language and Literacy Policy for Papua New Guinea
(National Literacy and Awareness Council, 1989 ) recommended the development of Tok Ples literacy programmes followed by an expansion of literacy skills in Tok Pisin, Hiri Motu and English. It suggested the following strategies:
1 Tok Ples preparatory classes for pre-school children;
2 Tok Ples bridging classes in Grade One of Community
Schools; 39
3 Tok Ples literacy maintenance in Community Schools; 4 Tok Ples instruction for non-core subjects in Community Schools;
5 Tok Ples maintenance programmes at the secondary and tertiary levels;
6 Tok Ples progammes for adults and out-of-school youth. This policy has been endorsed by the Secretary of Education and is now the policy which guides the development of literacy in Papua New Guinea.
Litteral (1975, p. 156 ) empha sises the importance of maintaining multilingualism in Papua New Guinea : Each language is used with specific functions in
specific situations so that the more languages a person knows , the more people he will be able to communicate
with .
Clyne (1982, p.27) uses Haugen 's term language ecology to express 'the wholesomeness for the human environment of the maintenance (and development ) of multilingualism'. Haugen
(1974 ) suggests that linguists should look beyond language itself and look at the total ecology, including such aspects as relation to other languages, demography, domains of use, other languages used with it, varieties , institutional
support and attitudes towards it.
For the multilingual situations which exist in most parts of
Papua New Guinea, it is probably not very useful to describe them in terms of the high-low distinctions made by domain
theory. Thurston (1992, p.124 ) points out that in 40 north-western New Britain, certain vernaculars which have been used as lingua francas, rather than replacing neighbouring languages, are now held in low esteem by their speakers. On the other hand, other vernaculars , with no value as lingua francas, have high loyalty amongst their speakers and are used to demonstrate ethnic distinctiveness.
In monolingual ecologies, variations are measured as dialects and varieties (registers and styles) of the language . In the language ecology of a multilingual society, code-switching is a common variation (McConvell,
1985) . Scotton (1982, 1988; Myers-Scotton , 1993a, 1993b, 1993c ) shows that intra-group code-switching encodes interpersonal dynamics and that, contrary to the beliefs of the domain theorists, "the possibility of code-switching is a motivation to maintain multilingualism in a community" (Scotton, 1982, p.440). Myers-Scotton (1993a, p.481) explains that "when speakers have social profiles encompassing the identities associated with two languages, and have the desire to signal these identities, then code-switching itself may become their unmarked choice". Such unmarked code-switching is prevalent in stable multilingual communities such as Singapore where an international language has official status but speakers have a different first language or speak an indigenous lingua
franca (Myers-Scotton, 1993a).
So, to maintain a language in a multilingual community, it is necessary to create and maintain harmonious relationships
among all the different codes in the speech community. And, 41 as Muhlhausler (1992, p.174) suggests, "a healthy linguistic ecology is one where dialectal variation abounds and where closely related languages can borrow from one another".
2.4 Social networks
Many sociolinguistic studies (e.g., Labov, 1972a; Trudgill, 1983; Hudson, 1980) indicate that extra-linguistic variables such as social class, situational context, age, gender, and region of origin affect language use . Milroy (1987a) adds the concept of 'social network ' to the list of variables.
The social network of an individual is the ser1es of personal relationships which that individual builds and basically comprises the people that the individual has come in contact with (Barnes, 1954; Bott , 1955, 1957). Some of the different types of contacts are members of one 's family, friends, workmates and those in voluntary associations such as fellow soccer players , fellow parishioners, and church ministers. The social network operates like a mechanism both for exchanging goods and services and for "imposing obligations and conferring corresponding rights upon its members" (Milroy, 1987a, p.47). It is the structural link that binds individual members together. Studies
investigating networks and language use include Gal (1979),
Milroy and Milroy (1985) and Bortoni-Ricardo (1985).
Schooling (1990) has also adapted network analysis and used it in his study of language choice in Vanuatu. 42
Milroy (1987a) defines an individual 's network as "simply the sum of relationships which he or she has contracted with others" . Gal (1979, p.l40) defines the social network as "all of the people an informant spoke to during a unit of time". The concept of 'network', according to Milroy (1987a), allows quantification of the character of an individual 's everyday social relationships - the influences to which he or she is exposed. The structure, content and diversity of individual network links can be both directly observed and quantified . The number of links they have with other individuals provides a measure of the density. The content of the link is given by the nature of the tie, e.g., kinship tie (husband or daughter ) or an occupational tie (workmate or boss ). The diversity of the link (viz., the number of ties which exist ) characterises the multiplexity of the link. Personal network structures of individual speakers are important in predicting language use (as shown by Milroy, 1987a) and language choice (as shown by Gal, 1979).
The social network, as a social category, is fairly
important in helping speakers choose appropriate varieties for presenting themselves linguistically in specific
situations. The language variety they choose reflects the type of network of which they are a part or with which they
wish to be identified (Le Page and Tabouret-Keller, 1985) .
As Gal (1979, p.15) says "social networks influence people 's
communicative strategies when group identification is expressed through speech". On these occasions, utterances
can be regarded as "acts of identity" (Le Page and 43
Tabouret-Keller, 1985, p.14) in which speakers "reveal both their personal identity and their search for social roles" . For some speakers, revealing their identity may not be easy, as Ludi found in a study of French-speaking migrants living in a German/Swiss German diglossic community - a migrant in a situation such as this is a "stage for identity conflicts" (Ludi, 1992, p.67)
Milroy (1987a) discovered that people with dense, multiplex networks in Belfast tend to stick to the vernacular variety of the language . This indicated identification with local norms within the three different communities she studied; namely, the Hammer, Clonard and Ballymacarett . Blom and Gumperz (1972) show the relationship between communicative strategies , identity and social networks in their study of language use in a Norwegian village. They recorded two groups of locals who were differentiated by their social networks . One group had more local contacts while the other group's networks consisted mainly of college students or outsiders as this group comprised college students who were home on holidays . Both groups agreed that the local dialect was appropriate for informal local discussions and it strongly symbolised pride and solidarity with the local
identity they felt . The standard dialect represented
national culture and the outside world.
The social network can be used to explain individual
behaviour of various kinds which cannot be accounted for �n terms of corporate group membership. But social network
theory can also link the way the network structures on one 44 hand, and the identity variables of age, sex, and area on the other hand can influence a speaker's language use. For example, Milroy (1987a) notes that in most of the literature on working-class communities gender is correlated with the density of the network. Women typically form denser and more multiplex networks .
Granovetter (1973, p.1360) argues that "it is through these networks that small-scale interaction becomes translated into large-scale patterns". He draws a distinction between 'strong ' and 'weak' ties, where the 'strength ' of the tie is defined in terms of time , emotional intensity, intimacy and the reciprocal nature of the tie . He then argues that the extent to which strong ties exist between individuals in a network will determine the density of that network (i.e., the extent to which all individuals will interact with each other ). Granovetter (1973) reasons that individuals with many weak ties are most likely to be the innovators in a community . Milroy and Milroy (1985, p.380) show that
"innovations are normally transmitted from one group to another by persons who have weak ties with both groups".
Milroy and Milroy (1992 ) use Granovetter 's concept of weak
tie (Granovetter, 1973) to link network analyis and social class-based analysis into an integrated model of
sociolinguistic structure . The social class-based analysis which they use is that of H¢jrup (1983), who regarded any
society as divisible into three different economically
determined 'life-modes ', each with different orientations to
work, leisure and family . Life-mode 1 is that of the 45 self-employed , for whom work , leisure and family are interlinked (e.g., in a family-run business ). Milroy and Milroy (1992, p.20) say that this life-mode will have "a close-knit type of network structure and a solidarity ethic". Life-mode 2 is that of the ordinary wage earner, for whom work is simply a means of earning an income and is not linked with leisure and family. This life-mode is very often associated with close-knit network structure and a solidarity ethic, but this can vary according to the economic circumstances of the community . Life-mode 3 is that of the professional or managerial employee , for whom work provides purpose in life and family and leisure are seen as supportive of this work. This group will form many loose ties of a professional nature "through which
innovations and influence may be transmitted" (Milroy and
Milroy, 1992, p.21).
Thus, the life-modes described above lead to social and cultural differences and to differences of network structure (Milroy and Milroy, 1992). And, as Gal (1979, p.17) points out : it is through their effects on the shape of social
networks ...• that macrosociological factors can
influence the language choices of speakers in everyday
interactions. 46
2.5 Patterns of variation and network structure
It is possible to measure the differences in individual network patterns (using a social network scale ) so as to identify the links between selected variables and the linguistic patterns of the speech. Milroy (1987a) notes that a crucial variable underlying any measure of personal network structure is degree of integration into a close-knit group. She constructed a 'network strength scale ' to measure the network patterns of individual speakers in terms of their density and multiplexity. The following conditions are those on which points (on a six-point scale from zero to five ) are scored:
1. being the member of a high density, territorially based cluster;
2. having substantial ties of kinship in the neighbourhood (more than one household in addition to his own nuclear
family ); 3. working at the same place of work as at least two
others from the same area; 4. having the same place of work as at least two others of the same sex from the area;
s. voluntarily associating with workmates in leisure
hours.
Statistical analysis of eight phonological variations and age, sex, area and network score were made . The results
confirmed her hypothesis that there was a correlation between a speaker 's heavy use of the vernacular and the
internal structure of his or her personal relationships. 47
That is, if an individual has close network ties with the local community, then his or her language variety will be close to the localised vernacular.
Using the social network concept Milroy (1987a) was able to examine the links of individual extralinguistic variables and language use. She was able to consider respondents individually through their networks and how they related to the local vernacular culture . Thus, personal network structures are shown to be valuable in predicting language use. Unlike some of the sociolinguistic variables (e.g., class or caste ) it can be applied universally.
In her study of the bilingual town of Oberwart in eastern Austria, Gal (1979 ) used Guttmann implicational scaling to display the language choices of her respondents and to rank the respondents according to their language choices. Gal says that displaying language choice in an implicational scale creates a predictive model in which conversational language switching is one type of variation within a larger framework of variation .
She (Gal, 1979) measured each person 's 'degree of
peasantness ' using the following criteria: Did the individual live in a household that :
( 1) owned cows;
( 2 ) owned pigs;
( 3) owned chickens ;
( 4) grew grain;
( 5) grew potatoes; 48
(6) used an outhouse or a flush toilet; (7) baked bread or bought it a store; Did the individual : (8) wear an apron (men) or a kerchief (women ) both indoors and outdoors regardless of weather;
(9) only do agricultural work as distinct from a regular job outside of agriculture;
10) attend only the eight mandatory years of school with no additional training in craft or commerce; (11) stop school after commercial or craft training, or go on to university or higher technical schools .
Gal found that the correlation between ranking of respondents according to their degree of peasantness and their ranking according to language choice was relatively low. She points out that even if this were not the case, a hypothesis would be required that would "outline how the status of the speakers influences the genesis and maintenance of differences in language choice" (Gal, 1979, p.140).
Gal (1979) also studied the statuses of people with whom the respondent often interacted (i.e., the members of the
respondent 's social network ) and ranked them on a peasant
to-urbanite continuum. She found a high correlation between
the ranking of respondents according to the degree of
peasantness of their social network and their ranking
according to language choice . Chapter 3
METHODOLOGY
3.1 Aims and Objectives of this study
This project was carried out as a study on language choice, 'language attitudes and social networks . Using questionnaires, interviews and observations, I set out to collect specific data showing which languages were reportedly used in certain situations e.g., home and work,
and reported reasons why this is the case . In addition, questions regarding respondents' views on language choice
were asked. Data regarding language used with different interlocutors were collected.
Although this is basic research, the results of the study may have important implications for an immediate practical
problem, i.e. , language maintenance .
3.2 Data collection
For an adequate explanation of linguistic change, Labov
(1972b, p.516) insists that one must analyse "data from the speech community", i.e., data on language use and attitudes,
together with linguistic data. This study is sociolinguistic in nature, mainly because it attempts to 50 explain language use in its social context i.e. , "where meanings are exchanged" (Halliday, 1978) , in the urban village of Hanuabada . There are numerous ways in which the linguistic and sociolinguistic information that is sought can be collected (Milroy 1987b). This study used questionnaires as a means of obtaining the information .
The responses which comprise the data of this study are the self-reports given by the people of Hanuabada. What is reported about use may not be an accurate reflection of linguistic reality when cross-checked with participant observations and recordings , but even these mis-reports may provide an insight into the language attitudes of the respondents.
The sample for the first part of the study was selected using stratified systematic sampling, with the clans as strata. The sample was proportionally allocated (Wiersma, 1986) based on the population data for Hanuabada obtained in the 1990 Census1 . I used systematic sampling by selecting every sixth house and then asking the first available and
consenting person in the house to be a respondent .
1 The clan population data are not published in the official census results, but I was given access to the unpublished data for the area/section of which Banuabada is a part. 51
The sample for the second part of the study was not proportionally allocated. It was designed to record respondents' social networks to look for possible correlations with language choice . An attempt was made to stratify the sample using optimum allocation, i.e., to ensure that a variety of social networks and language use patterns were represented. Thus, something that I deliberately attempted to do in the second part of the study was to identify the networks of people who lived out of the village and especially in the urbanised parts of Port Moresby. I thought this sub-group of respondents might have a crucial role in diversifying the language situation. I saw their social networks operating differently from those of respondents who were based in the village . To identify and then evaluate the strengths and weaknesses of the different types of networks might well be a step towards understanding language maintenance , if language choice is
determined by the type of network contacts that a speaker
has .
The first part of the study was designed specifically to investigate the language use and attitudes of people in
Banuabada . The second part was designed specifically to
investigate the "villageness'' of the respondents and their social networks , i.e. , the number of people they came in
contact with most often and how the respondents related to each of them. As with the respondent , the villageness of
each of his or her contacts was also recorded. 52
I used the following five indices to measure the degree of villageness: (1) where the respondent/contact spent his/her early life;
(2) where the respondent/contact resides ; (3) where the respondent/contact spends the working hours of the day (whether working or not); (4) whether the respondent/contact spends more than fifty per cent of his or her free time in the village; and,
(5) whether the respondent/contact attends village ceremonies, e.g., bride-price and funeral feast ceremonies. One point was allocated to each feature if the response implied a strong connection with the village (e.g. , one
point was given if the respondent spent his/her early life in the village, but no point was given if the respondent
worked in the city) . This gave a maximum of five points for "complete villageness". The "actual villageness" was then
expressed as a percentage of the complete villageness (i.e.,
where five points = 100 per cent villagness).
On arrival in Hanuabada , I decided to take a count of all
the houses in Hanuabada . I did all the counting in person with a little help from my sisters. I also managed to find
out the name of the head of each household as well as the clan to which each belongs . I thought that if I had
difficulty obtaining census data for a stratified random
sample using proportional allocation, using the number of
houses per clan from my own counting would suffice. In the
end there was no problem having access to the pre-published
census data so I used the census information. However, the 53
1992 house count did provide an interesting comparison with the census data. Tables 3.1, 3.2 and 3.3 contain information that has been derived from the house count and
the 1990 census data.
The term "outsiders" refers to people who have migrated to the village and settled there . Some have migrated from other places like Rigo, which is about 100 or more kilometres south east of Port Moresby, or Goilala in the highlands area of the central province of PNG. Some outsiders are from Kerema in the Gulf province. The Kerema
people used to be the traditional trading partners of the
Hotuans in the famous Hiri trade voyages which are now
defunct. Some of the outsiders from Kerema would have come to Banuabada on those trips and decided to stay, making Banuabada their home . Informant number 3 fits this category. His grandparents came to Hanuabada on one of
those Biri trips and settled there. He and his immediate family regard themselves as Hanuabadans. This family has acquired the language and the lifestyle, and affiliated to the Havara clan of Banuabada. However, because they do not
originally come from Banuabada , I have grouped them as outsiders for my research purposes.
Other outsiders come from the other section of Poreporena
village - Elevala. I found several heads of households from Elevala who have built houses on the land space of Hanuabada and happily live there . In other words, there are Elevala
people migrating to Hanuabada and building their homes
there. 54
Generally, potential migrants have to be introduced to the village by a Hanuabadan person or family either as family friends, or friends of close friends, or as adopted families. Once they have settled and established themselves without causing any serious problem or harm to anyone in the village in the first few months , the migrants are welcome to stay. But to the villagers of Banuabada (and also for this study) these migrants and their children are classed as outsiders . However, the children of mixed marriages where one partner is of Banuabada (Motu-Koita) origin, would be classed as Banuabadan .
As can be seen in Table 3.1, there are 364 houses in the village . Host of them (73) belong to the Gunina clan. The
Kahanamona clan has the smallest number with only eight houses (just over two per cent of the total). Note that just over nine per cent of the houses are occupied by outsiders . 55
Table 3.1: 1990 Census figures for Hanuabada
Clan Douses I n h a b i t a n t s No .J No. % male female total % house
Gun ina 73 20.05 393 393 786 19 .53 10 Bohodae 52 14 .29 287 278 565 14 .04 10 Apau/Vahoi 49 13 .46 302 257 559 13.89 11 Botai 45 12 .36 300 254 554 13.77 12 Tubumaga 42 11.54 226 234 460 11.43 10 Outsiders 33 9.07 148 132 280 6.96 8 Mavara 32 8.79 185 172 357 8.87 11 Kwaradubuna 30 8.24 192 180 372 9.24 12 Kahanamona 8 2.20 53 38 91 2.26 11
TOTAL 364 100.00 2086 1938 4024 100.00
Table 3.1 also shows there were 2086 males and 1938 females in Banuabada in 1990. The number of people per house averages out to 11, but the outsider households have a smaller average of eight inhabitants per house . Ninety-one people (just over two per cent of the village population) belong to the Kahanamona clan whilst 786 people (nearly 20 per cent of the population) belong to the Gunina clan. The census also found that 280 people (nearly seven per cent) inhabited the 33 houses headed by outsiders. The Mavara clan and the Kwaradubuna clan each have just under ten per cent, whilst the Apau/Vahoi, Botai, Hohodae and Tubumaga clans each have between 11 and 14 per cent of the population. 56
I counted 22 outsider houses while the census counted 33, making a discrepancy of 11 houses . As can be seen �n Table 3.2, there are discrepancies in the data for all the clans that are listed.
Table 3.2: 1990 Census figures compared with 1992 count
B o u s e s People Clan 1990 Census 1992 Count 1990 Census No. % No . % %
Gun ina 73 20.05 71 19 .24 19 .54 Bohodae 52 14 .29 49 13.28 14 .04 Apau/Vahoi 49 13.46 59 15.99 13.89 Botai 45 12 .36 55 14 .91 13.77 Tubumaga 42 11.54 38 10.30 11.43 Outsiders 33 9.07 22 5.96 6.96 Mavara 32 8.79 25 6.78 8.87 Kwaradubuna 30 8.24 39 10.57 9.24 Kahanamona 8 2.20 11 2.98 2.26
TOTAL 364 100.00 369 100.00 100.00
I collected data on two separate occasions when I was in Port Moresby . I asked people if they were willing to participate in the study that I was conducting for my post-graduate studies. If they answered "yes" (which was �n most cases ), I then interviewed them using the questionnaires in Appendices 1 and 7. The questionnaires are written in English, but for those respondents who preferred it, I asked the questions in Motu . Some preferred to take the paper away and I collected them a few days later when they were completed . Some decided to complete the 57 questionnaire there and then without me interviewing , although they consulted me if any difficulties arose throughout the answering process.
The questionnaires were administered in the subjects• natural surroundings which in some cases were residences, and communal outdoor recreation areas in other cases . The location was chosen largely by the participants themselves as I thought it very important for them to feel at ease .
The interview sessions were not consistent in their length; some took hours and others half a day or even a full day. The reason for this is because, as an insider, I had to observe the cultural norms relating to conversation. For example, this involved allowing the older respondents a lot more time to answer the questions . Respondent number 6, a man aged 86 years, wished to give me the history of the Bohodae clan (unfortunately this was not recorded ) and this took most of the day . For many of the sessions I was asked to stay and eat a meal. Because of the cultural norms , even though I was eating their hard-earned food, acceptance was a form of appreciation of their offer and a "bonding,. to the group. But this acceptance in each case did make the interviews take more time .
During the first visit, the first five days were leading up to the national elections and I was unable to conduct interviews as people were preoccupied with election issues. I was staying in Gordons , about five miles from Hanuabada, and I would travel to the village at 9 a.m. and return home 58 at 5 p.m. each day . Lack of transport made it difficult to travel on some days - and in all, 11 days were spent collecting data for Questionnaire 1.
The second visit was for a period of six weeks . During this visit 32 days were spent collecting data for Questionnaires 1 and 2. As an insider, I felt welcomed and on some occasions I was noted as a personal contact on the respondent 's questionnaire sheet . Other than the three people who declined when I approached them, all those approached willingly participated in the study.
Of the forty people who agreed to participate during my first visit , 19 were males and 21 females. In the second round, I managed to interview 15 more people from selected houses which had not been visited on the first round . Thus, for Questionnaire 1, I ended up with 55 people in the sample: 30 males and 25 females.
Over one third of the interviews were recorded. In addition, approximately seven hours of spontaneous conversation was recorded together with two church services.
All the respondents are associated with Hanuabada in one or more of three different ways: a) they are originally from Hanuabada ; b) they are married to someone from Hanuabada and reside
in the village; 59 c) they are originally from a different area, but as a result of their ancestors migrating to Hanuabada they were born and raised in Hanuabada, and therefore regard themselves as members of the Hanuabada community.
As Table 3.3 shows, my sample for Questionnaire 1 is proportionally allocated according to the percentage of the inhabitants of the houses in each clan (based on the 1990 census figures ). In this way , the clan distribution in the sample is quite representative of Hanuabada .
Table 3.3: Sample 1 and 1990 Census population figures
Clan Census Sample 1 Variation % No . % %
Gun ina 19 .53 11 20.00 .47 Bohodae 14 .04 8 14 .55 .so Apau/Vahoi 13 .89 8 14.55 .65 Botai 13.77 7 12 .73 1.04 Tubumaga 11.43 5 9.09 2.34 Kwaradubuna 9.24 5 9.09 .15 Mavara 8.87 5 9.09 .22 Outsiders 6.96 4 7.27 .31 Kahanamona 2.26 2 3.64 1. 37
Total 100.00 55 100.00 7.07
As can be seen, n1ne Hanuabada clans are represented, although the Apau and Vahoi clans are shown as one because of their closeness in the community and the small numbers in the sample. The Outsiders group includes respondents from 60
Ewau (a Misima clan in Milne Bay Province), a clan from the Sepik province and one from Kerema in the Gulf province.
Table 3.4: Sample 1 and 1990 Census age group figures
National Capital Sample 1 Age District Respondents Group No . % No. %
0-12 66703 34 .11 0 .00 13-20 36962 18.90 7 12 .73
21-35 60210 30.79 18 32 .73 36-50 24019 12.28 20 36.36 51+ 7676 3.92 10 18.18
Total 195570 100.00 55 100.00
As can be seen in Table 3.4, the census figures (National
Census, 1990) show us that the National Capital District has a young population. Thirty-four per cent is under 13 years old. Fifty per cent is between 13-35 years old and only about four per cent is over 51 years. No attempt was made to make the age distribution of my sample representative. A comparison of the age distribution for sample 1 and the age distribution for the National Captial District is given in
Table 3.4. 61 .
The male/female distribution for sample 1 can be seen in Figure 3 .1.
nale
re ..alo • � 1: 28 II � 1: 0 &:a. 15 • � "" 0 18 Cl til • � 5 1: II u " II 8 A.
Ll-28 21-35 36-58 51•
Age Group
Figure 3.1: Age groups to which respondents belong
In the second visit , I interviewed 52 people for Questionnaire 2: 29 males and 23 females. Because the main purpose of the second part of the study was to investigate the relationships between language use and social networks , it was not considered important to ensure proportional allocation in accordance with the percentage of the inhabitants of the houses in each clan. Instead, it was important to ensure that participants had social networks 62 across the whole of the villageness and language-use ranges . Thus, as with other studies of networks , the sample design provided (1) less basis for using statistical inference to generalise the research results to a larger population, but (2) a greater capacity to understand the nature of communication structure. (Rogers and Kincaid, 1981:102-3 )
Table 3.5 shows my sample for Questionnaire 2 in comparison with the 1990 Census figures.
Table 3.5: Sample 2 and 1990 Census population figures
Clan Census Sample 2 % No. %
Gun ina 19.53 10 19.23 Bohodae 14 .04 7 13.46 Apau/Vahoi 13.89 5 9.62 Botai 13.77 9 17.31 Tubumaga 11.43 2 3.85 Kwaradubuna 9.24 5 9.62 Mavara 8.87 7 13.46 Outsiders 6.96 1 1. 92 Kahanamona 2.26 6 11.54
Total 100.00 52 100.00
In order to determine language use, Questionaire 2 asks respondents which languages they use with particular interlocutor types . In this context, an interlocutor is any 63 person with whom the respondent interacts on a fairly regular basis. The relationship which exists between the respondent and the interlocutor (which may be personal, social , spiritual, or other) can be used to classify the interlocutors into types . Examples of these types are : grandparents, friends , employers , doctors, teachers , parish priests and shop assistants. Page 2 of the questionnaire (see Appendix 7) shows the types I have researched, which were chosen so as to cover a range of domains of interactions.
3.3 Data analysis
The answers provided by the respondent were coded and recorded in the tables shown in Appendices 2-9 . For those questions with open-ended responses, the answers were subjected to content-analysis (the answers were categorised according to the type of content matter ) and the resulting content-statements were assigned to coded categories. For example, the content-statements about how literacy in a particular language was acquired were assigned to just three categories. (See the key to Appendix 2.) The answers to questions 12 and 13 were sufficiently similar to be combined in the analysis of the results . Similarly, the answers to questions 16 and 17 were combined . The data for each question were analysed and the results displayed in the form of bar charts, pie charts and tables. 64
The information relating to language choices with different interlocutor-types was arranged in a matrix, with the interlocutor-types as the columns and the respondents as the rows . The respondent's language choice was recorded as M (for Motu ), 0 (for Other) or MO (for Motu and Other) in each cell. The rows and columns were then rearranged (using a computer spreadsheet ) so as to form a "'scalogram', which at a glance gives the configuration of the qualitative data" (Guttman , 1944, p.l39).
The aim of the rearrangement of the columns was to produce a configuration such that across each row: only an M appears to left of an M; only an 0 appears to the right of an 0; only an MO or an M appears to the left of an MO; and, only an MO or an 0 appears to the right of an MO. In this way, the interlocutor-types were ranked to form an "implicational scale" (Guttman, 1944).
In the same way , the aim of the rearrangement of the rows was to produce a configuration such that in each column: only an M appears above an M; only an 0 appears below an 0; only an MO or an H appears above an MO; and, only an MO or an 0 appears below an MO. In this way , the respondents were ranked to form an implicational scale showing the pattern of variation between them as speakers.
A perfect scalogram would be one in which every cell satisfied the conditions for columns and rows above , and so enabled predictions to be made with 100 per cent accuracy . 65
The degree of approximation to perfection is measured by a coefficient of reproducibility, which is the empirical relative frequency with which values of the attributes do correspond to intervals of a scale variable. (Guttman , 1944, p.150) The co efficient of reproducibility (CR) for the scalogram produced in this study was calculated by the following formula: CR = 1 - number of errors number of questions x number of respondents The coefficient of reproducibility was then expressed as a percentage to give the "scalability" of the scalogram. Guttman (1944) regards 85 per cent as the minimum scalability required for the scalogram to be classed as "scalable".
After achieving maximum scalability, some of the rows and columns were then rearrranged (without reducing the scalability ) to bring the errors as close as possible to the boundaries or "cut points". The resulting "best-fit" language-choice scalogram was then used to produce a rank order for the respondents. The "villageness" of each respondent 's social network was calculated and this index was then used to rank order the respondents. The Spearman rank order correlation coefficient was used to measure the correlation between the two ranks and hence to measure the association for the variables 'social network villageness ' and 'language choice '. 66
3.4 Ethical concerns
As with all studies involving people, it was important to ensure that the study conformed to appropriate ethical standards. Before conducting any interviews or recording speech, I sought the consent of the individual to be studied . Apart from the very few who declined, the majority of respondents willingly agreed to participate . Although some participants didn't mind being named, I explained that they represented a group and that it was more ethical for me to refer to them as anonymous members of that particular group under study . Participants were simply referred to by a number in the analysis and in my reporting of the findings - this anonymity ensures that individual participant confidentiality is preserved.
I began each encounter with a brief explanation of why I was
carrying out the exercise; that is, why I was collecting the information, why I needed it and how I was going to use it . I also made the participants aware of my long term ideas and objectives. I felt that it was important that the participants were very clear about what they were getting involved in and why. Any questions and queries were satisfactorily answered before participants committed themselves. 67
3.5 Limitations of the study
This is a cross-sectional study for which the data were collected at one point in time . Therefore, since this is not a longitudinal study, one can only infer processes of language shift.
For the first part of the study, the sample was stratified with clans as the strata . It is clear to the casual observer that in Banuabada there are many children, and since no children were included in the sample it is clearly not representative in this respect . Age was not used as the strata and consequently the sample is not representative of
the age profile of the population . Instead, I felt that it was important to include a sufficient number of people in each of the categories so as to ensure a range of views on language and cultural maintenance issues.
Of course, the languages that children use daily in Banuabada may be the most significant factor in language maintenance. It would have been very useful to have collected some views from the children themselves. However, the responses received from the very few respondents under the age of 13 years did not prove constructive or helpful and therefore were not included in the sample. Shyness was the major hindrance, although there was one case where the mother 's interferance dominated the interview. Perhaps the children from Banuabada were not prepared mentally and socially for such an exercise in the more or less formal 68 atmosphere of an interview. Maybe an interactive group interview in a casual, informal style with children of very similar ages would be more productive.
My "insider" status may have limited the study in that people may have been inclined to give me the answers which they thought I wanted to hear or which they wanted me to hear. This is a version of the Observer 's Paradox , i.e., the phenomenon whereby the presence of the observer affects the knowledge that the investigator ultimately acquires of the linguistic repertoire of the group under study, i.e., the type of data collected for analy sis. Also, as an insider it is sometimes difficult to observe and interpret behaviour in an unbiased way . However , both these phenomena are also problems to some extent for "outsider" researchers .
Obviously, an investigator must acqu�re knowledge of the local norms and values and understand them sufficiently to be able to obtain and analyse the data. For a stranger entering an alien society, a knowledge of what to say is not enough. Be or she also needs a knowledge of what type of codes, channels and expressions to use in what kinds of situations to what kind of people (Hymes, 1962, 1972).
Community interactional patterns cannot always be understood without a long period of observation to establish the local norms and values of the community, which may be quite different from more publicly accepted norms and values
(Milroy, 1987a) . As an insider, my knowledge of the interactional patterns of the community enabled me to conduct my research in a culturally acceptable manner and to 69 analyse the data using this knowledge . Also, my knowledge of the languages spoken by the respondents enabled me to translate the English questions when required to do so .
The data collected on language use in this study are based on the self-reports of the respondents. These self-reports may not reflect actual usage . According to Milroy (1987a), the most coherent models available for investigation into the study of a language in its community setting are those of Labov (1972b) and Blom and Gumperz (1972). For example, recording group interactions allows the investigator to have access to speech in natural and informal situations (Labov, Cohen, Robins and Lewis, 1968; Labov 1972b) . These recordings can then supplemented by data collected from interviews. I did record group interactions but was unable to include an analysis of them in this thesis.
Part 1 of the study looks at language use in the home , in school, at work, and in the city . However, since the survey instrument did not consider language use in relation to the interlocutors or the topics, then the data collected relate to language use in these situational contexts rather than to domains . Chapter 4
MAIN RESEARCH FINDINGS - PART 1
LANGUAGE USB AND ATTITUDES
4.1 Importance of Motu and other languages
As can be seen in Appendix 4, 83 per cent of respondents reported that Motu was very important and 15 per cent said it is fairly important . Thirteen respondents (24 per cent ) gave cultural maintenance as their reason for regarding Motu as very important whilst 56 per cent said that it was because it was their mother tongue .
Nearly three quarters (72.5 per cent ) of the respondents said that English was just as important as Motu. Twenty four per cent said that English was important because it is a lingua franca , 16.4 per cent said that its importance could be attributed to its use as an international or universal language, and 12.7 per cent believed it was important because it is the medium of instruction in Papua
New Guinea.
Tok Pisin was said to be equally as important as Motu according to 12 .7 per cent of the respondents, who all gave as the reason - "because it is a lingua franca". 71 .
4.2 Languages spoken at home
Figure 4.1 shows that 64 per cent of the respondents use only Motu in the home , 31 per cent use Motu and English, two per cent use Motu and other languages , and the remaining three per cent use English alone or with other languages .
Thus, 96 per cent of the respondents use Motu in the home , solely or in combination with some other language - making Motu the preferred language in this domain .
D """' • Eng (J!..;<) • ""tu • Other (�) 0 Eng • Other ()".<)
Figure 4.1: Languages spoken at home
Respondents 21 and 55 reported English and another language as the home language as a result of having parents from different races, and a spouse being non-Motuan. 72
Figure 4.2 shows the reasons which were given to explain why Motu is used in the home . Twenty-nine per cent of the sample use it because it is spoken by everybody else in the
horne and 13 per cent use it because it is spoken by everybody else in the village . Fifty-eight per cent use it because it is the mother tongue and two per cent gave reasons related to culture and language maintenance .
78
• .. c 68 • 1 58 •l � ... 48 0 • • • 38 .. c 28 • ..� 18 c==J 8 11T H v u c
Reasons
Key:
MT • .ather tongue U • helps understanding
H • spoken In the hoae C • keeps culture and language strong
v • spoken In the vi l lage I • improvement of the language
Figure 4.2: Reasons for Motu being used at home 73 -
As was shown in Figure 4.1, English is used at least some of the time in 35 per cent of the homes. Figure 4.3 shows the reasons given by respondents.
14 12 18 B • • G • 4 1• 2 .. 8 �-- ��--��-- X A I s F
Reasons l(�y: X • p•r�ntfspous� sp�•ks •nother longuoge
A • us�d wh�n non-Hotuons ar� present I : to improve the language
S • spoken by ev�ryon� �lse in the home F : first language
Figure 4.3: Reasons for English being used at home
Twenty-two per cent of the fifty-five respondents use English because of the presence of non-Motuans as close family members (spouse/parent ) or as other relatives or visitors . Ten per cent use English in order to assist with English language development or to assist at work and school. 74
In expressing their views on which languages should be
'learned at home• , 67.5 per cent said "Motu" . As can be seen in Table 4.1, over half of the respondents (57.5 per cent ) felt that it should be learned at home because it is their "mother tongue".
Table 4.1: Reasons for learning Motu at home
Reason given: %
CL communication link with fellow Motuans 7.5 MT mother tongue or native language 57 .5 U helps understanding 2.5
NB : some respondents gave two reasons
Table 4.2 sets out the reasons given by the 72.5 per cent of respondents who said that English should be learned at home . Twenty per cent of respondents reported that English should be learned at home because it is the medium of instruction and 40 per cent because it helped communication at school and work.
Table 4.2: Reasons for learning English at home
Reason given: %
A for when non-Motuans are present 2.5 IU international/universal language 9.1 LF communication link with other groups 2.5 M medium of instruction 20.0 S for communication at school or work 40.0
NB: some respondents gave two reasons 75 -
4.3 Languages used at school
As can be seen in Appendix 2, all the respondents had
attended or were attending school. Nearly 13 per cent of
the respondents were women who had achieved grade 10, and 22
per cent were men who had achieved the same grade . Overall ,
42 per cent reported that they had reached grade 10 or
over. It should be noted that none of those who were over
the age of 50 had achieved levels of education beyond mid or
upper primary . Figure 4.4 shows the distribution of grades
achieved by the respondents.
Fe••le 0 n.le
25
• .. Zit � l
1• 15 .I .. 0 • • d ..• c ! l s
•
3 .. s 6 7 0 9 18 11 12
Grade Ach ieved
Figure 4.4: School grades achieved by responden ts 76
Figure 4.5 represents the languages used in the school classroom as reported by the respondents.
Eng (5Sz)
�tu • Eng (3�/.)
llotu
Figure 4.5: Languages used at school
Ninety-five per cent used English at school, the majority
(82 per cent ) reporting that this was because it was the language of instruction . Five per cent reported Motu only as the school language . This group belonged to the 50+ age category . They reported having attended schools that were conducted 1n Motu by the LMS missionaries, although English was used for writing and in the subsequent higher standards . Appendix 4 shows that a higher percentage (85 per cent ) of respondents favour learning English, rather than Motu (10 per cent ), at school . 77
Table 4.3 shows the respondents 's views on the Tok Ples education policy proposed by the Papua New Guinea Education Department .
In response to the statement that there should be Tok Ples pre-schools in Banuabada, 27 per cent of the respondents agreed but 71 per cent disagreed. Seventy-seven per cent of the respondents disagreed with the statement that there should be Tok Ples bridging classes in Grade 1, and 67 per cent disagreed that Tok Ples literacy should be maintained in community school .
Table 4.3: Attitudes to Tok Ples education
Views on Tok Plea education: Agree Uncertain Disagree % % %
There should be Tok Ples
pre-schools in Banuabada ••••. 26.92 1.92 71.16 There should be Tok Ples bridging classes in Grade 1 .. 19.23 3.85 76.92 Tok Ples literacy should be maintained in Community School 21.15 11.54 67 .31 Tok Ples should be used
in non-core subjects ••••••••• 17.31 17 .31 65.38 in Community School There should be Tok Ples maintenance programmes at the 23.08 15 .38 61.54 secondary and tertiary levels
There should be Tok Ples
programmes for adults and •••• 2 1 . 15 9.62 69.23 out-of-school youth 78
4.4 Languages used at work
As can be seen in the table in Appendix 2, 87 per cent of the respondents had some member or members of the extended family earning an income ; the remaining 13 per cent did not indicate that anyone in the family was employed . In 27 per cent of the responses, the respondents indicated that they themselves were employed in the formal sector. Twenty-three per cent of the respondents had a spouse who was employed (husbands , 18 per cent; wives, 5 per cent). Nearly 15 per cent had an employed sister and 12 per cent an employed mother.
Figure 4.6 shows the languages and combinations of languages used at work (N=44). Twenty-nine per cent use three or more languages at work . Ninety-five per cent use English at work : 68 per cent in combination with other languages and
27 per cent English only . 79
liD Ens (27"/.) Ill !btu • Eng (16xl OOl Eng • TP CZlxl • Eng • !btu • TP .dCZ?-
• Eng + TP • other CZXJ � !btu Figure 4.6: Combinations of languages used at work As can be seen in Figure 4.6, 48 per cent use Motu at work - 43 per cent use Motu in combination with other languages and five per cent use Motu only. It could be the case that for some respondents "Motu" was used to mean "Hiri Motu. 4.5 Languages spoken in the city Fifty people responded to the question about languages used in the city . Thirty-one per cent reported using three or more languages in the city . Seventy-eight per cent reported that they used Motu at least some of the time (i.e. , either 8 0 singly or in combination with other languages) - though, as mentioned above, some respondents may have meant "Hiri Motu ". Eighty-eight per cent reported that they used English and 48 per cent reported that they used Tok Pisin at least some of the time . As can be seen in the summary in Appendix s, most (90 per cent) of those who use Motu do so because it is the "common language" . Sixty-three per cent of those who use English and 71 per cent of those who use Tok Pisin do so because they are lingua francas . 4.6 Languages spoken in church Religion �s taken very seriously in Hanuabada as can be seen from the results shown in Appendix 2. Seventy-three per cent of the participants belong to the United Church (which used to be the London Missionary Society ). Twenty-five per cent are Catholics and two per cent belong to other church groups . Motu is the language used by the United Church, as indicated by respondents 6, 12 and 23. Respondent 12 indicated that Motu was 'very important ' because it was the language used in church services. 81 4.7 Literacy Table 4.4 shows how literacy was acquired in English, Motu, Tok Pisin and other languages. Ninety-six per cent reported that they had acquired English literacy at school ; five per cent of the respondents said that they had acquired English literacy while at work. Thirteen per cent taught themselves literacy in Tok Pisin and 65 per cent in Motu. Table 4.4: How literacy was aquired Bow literacy Responses for each language was aquired: Motu Eng TP Other % % % % ST self-taught 65 13 4 T taught at school 27 96 4 2 w learned at work 5 5 {NB: some respondents gave two explanations) Chapter 5 MAIN RESEARCH FINDINGS - PART 2 SOCIAL NETWORKS 5.1 Language choices Appendix 9 shows the languages which each respondent reportedly uses with the various interlocutor types selected for this study. Forty two (81 per cent ) of the respondents reported using at least three languages with this group of interlocutor types; nine (17 per cent) reported using two languages (Motu and English) ; and one reported using Motu only . Thirty seven (71 per cent ) of the respondents reported using at least three languages with one or more of the individual interlocutor types. Thirty one (60 per cent ) of the respondents reported using Motu , English and Tok Pisin and one reported using Motu , English and Other, with the interlocutor type 'friends '. Table 5.1 shows the scalogram produced by ranking the respondents and the interlocutor types to form an implicational scale, scaled with respect to the respondents' choice between Motu (M) , Other (0) or mixture of Motu and Other (MO). A dash ("-") indicates that the respondent did not report any language for this interlocutor type . 83 The scalogram reveals a clear pattern of language choice: from low use of Motu and high use of Other languages to high use of Motu and low use of Other languages. Thus Respondent number 6, who is ranked in position 1, uses Motu on its own only with Interlocutor 1 (grandparents) and uses mixed Motu/Other or just Other languages with the other Interlocutor types . At the other end of the implicational scale is Respondent number 30, ranked in position 52, who uses Motu on its own with all the Interlocutor types with whom she interacts . The horizontal and vertical lines show the 'cut points ' between the different choices, i.e., from Motu to Motu/Other to Other. The scalability for the matrix is 90.2 per cent, which meets Guttman 's criterion of 85 per cent to be scalable (Guttman, 1944 ). As can be seen, Respondents 1, 32 , 35, 40, 47 and 51 had four or more 'errors ', i.e., language choices which were not of the particular category of the choice zones which they occupied (these errors are circled on Table 5.1). Thus, for these respondents the scalability of their responses was less than 80 per cent, i.e., not scalable by Guttman 's criterion . Al so, columns 8, 12 and 16 (corresponding to interlocutor types 'neighbour •, 'teacher ', and •town shop assistant ' respectively) had ten or more 'errors '. Thus, for these interlocutor types the scalability was less than 82 per cent, i.e., not scalable by Guttman 's criterion. 84 Table 5.1: Scal ogram of language choices Resp Language Used With Each Interlocutor Type No . Rnk 1 13 9 14 2 3 21 8 4 6 5 7 16 17 12 18 24 23 22 06 1 M MO MO MO MO 0 MO MO - MO - 32 2 - 0 0 M M M M MO MO MO MO @X@) - 36 3 M M M M M M - @MO g� 09 4 0 0 0 0 0 0 M M M M M M MO 0 0 0 0 0 0 13 5 M M M M M M MO MO - MO OOMO �0 0 0 01 6 M M M M M MO MO MO MO - MO 0@)0 0 0 0 MO 02 7 M M M M M MO M - MO 0 0 MO 0 08 8 M M M MO M M MO MO M MO - MO 0 0 0 0 0 0 0 24 9 M 0 0 M M MO 0 M MO M MO 0 0 0 0 22 10 M M M M M M MO MO MO M - MO MO 0 0 0 45 11 M M MO M MO 0 MO MO MO MO MO 0 0 0 � 35 12 M M M M MO MO ®Mo - MO MO - 0 0 0 0 48 13 M M M MO =� MO MO MO MO MO MO MO - 0� 0 0 52 14 M M M M M M 0 MO MO 0 0 0 0 MO 11 15 M M M M M M M MO MO - MO MO MO 0 0 - 0 0 0 0 0 19 16 M M M M M M M MO MO MO MO MO MO MO 0 0 0 29 17 M M M M M MO M MO MO MO MO MO 0 MO 0 31 18 M M M M M M 0 MO MO MO MO MO MO - 0 07 19 M M M M M M M M MO MO so 20 M M M M M M M MO 0 0 0 0 � 0 0 0 51 21 M M M M M - 0 0 0 0 0 0 M M MO 0 0 0 0 46 22 � M M M M M M MO M MO 0 0 43 23 M M M M M M M - MO 0 0 0 0 - 0 0 0 0 0 0 0 44 24 M M M MO M M M M MO 0 0 0 0 0 0 0 49 25 M M M M M M M M M - MO MO 0 MO 0 MO MO 0 38 26 M M M M M M M MO MO MO � 0 0 0 0 47 27 M M -@ M :at� M - MO MO 33 28 M M M M M M MO M M MO 0 0 0 0 - 0 0 0 g� 0 12 29 M M M M M M MO 0 M MO MO 0 - 0 0 0 0 0 10 30 M M M M M M MO MO 0 0 0 0 0 0 40 31 M M �@ M M M M MO MO 0 0 28 32 M M M M M :� M M M MO 0 MO MO 0 ®® 18 33 M M M M M M M M M M M MO MO MO M M 04 34 M M M M M MO M M M - MO MO MO 0 MO 39 35 M M M M M M M 0 0 - MO 0 MO MO MO OS 36 M M M M M M M M M MO MO MO MO MO MO 03 37 MO M M MO M MO M - MO MO MO MO MO 27 38 M M M M M M M M 0 MO - MO MO MO MO MO 42 39 M M MO M M M M M M - MO MO MO MO MO 41 40 M M MO M M M M M M M - MO MO MO MO MO 15 41 M M M M M M M M M M - MO MO MO MO MO 17 42 M M M M M M M M M - MO MO MO MO MO 16 43 M M M M M M M M M M - MO MO MO MO MO 34 44 M M M M M M M M M M - MO MO MO MO MO 21 45 M M M M M M M M M M M MO MO MO MO MO 20 46 M M M M M M M M M M M MO MO MO MO MO 37 47 M M M M M M M M M M M MO MO MO MO MO 23 48 M M M M M M M M M M M MO MO MO MO MO 26 49 MO M M M M M M M M M MO MO 0 MO 25 50 M M M M M M M M M M M MO MO MO 14 51 M M M M M M M M M M M M MO MO MO MO 30 52 M M M M M M M M M M M M M M M M Scalability = 90.2% 85 Key to Table 5.1 Htmber Interlocutor Number Interlocutor 1 grandparents 13 village seller 2 parents 14 trade store assistant 3 siblings 16 town sho� assistant 4 spouse 17 bank ass�stant 5 grandchildren 18 government official 6 own children 21 fellow . worshipper 7 friends 22 superv�sor 8 neighbours 23 subordinate 9 traditional doctor 24 co-worker 12 teachers = no language reported M = Motu 0 = Other language MO = Motu + Other language Bach horizontal line represents a variation between speakers . Thus, the scalogram shows there to be 16 types of language users in the sample. In order to permit statistical testing and some degree of generalisation, this number of types needs to be reduced. This can be achieved by combining each type with five or less members with the neighbouring group which is most similar. By 'smoothing out • the minor deviations in language choices (represented by the shorter horizontal and vertical lines in Table 5.1), the scalogram of language use can be divided so as to give three broad types of language users, as shown in Table 5.3: 'low-rank language type ', which falls between the ranks 1-13 and comprises respondents who use Motu only with 32 per cent of the interlocutors , use Other with 36 per cent, and mixed Motu/Other with 32 per cent; 86 'mid-rank language type ', which falls between ranks 14-35 and comprises respondents who use Motu only with 58 per cent of the interlocutors, use Other with 32 per cent, and mixed Motu/Other with 11 per cent; 'high-rank language type ', which falls between ranks 35-52 and comprises respondents who use Motu only with 58 per cent of the interlocutors, do not use Other at all, and use mixed Motu/Other with 26 per cent . Table 5.2 shows the percentage of respondents in each language type and table 5.3 shows the configuration of these types in the scalogram. Table 5.2: Language ranks by respondents Language Respondents Ranks % Low (1-13) 25.00 Mid (14-34 ) 40.38 High (35-52 ) 34 .62 Totals 100.00 {N) (52) 87 - Tabl e 5.3: Scalogram showing language types Resp Language Used With Each Interlocutor Type No. Rnk 1 13 9 14 2 3 21 8 4 6 5 7 16 17 12 18 24 23 22 06 1 M MO MO - - - MO MO 0 MO MO MO 0 0 0 0 - - - 32 2 M M - - - M M MO MO MO MO 0 0 MO MO MO 0 MO - 36 3 M M - - M M M M 0 0 0 MO 0 0 0 MO 0 0 0 09 4 M M M - M M M MO 0 M MO 0 0 0 0 0 0 0 0 13 5 M M M M - - - M M MO MO MO 0 0 MO 0 0 0 0 01 6 M M - M - M M MO MO MO MO MO 0 MO 0 MO MO MO MO 02 7 M M - - - - M M M MO M MO 0 0 MO 0 0 0 0 08 8 M M M MO - M M MO MO M MO MO 0 0 0 0 0 0 0 24 9 M - 0 0 M M MO 0 M MO M MO 0 0 0 0 - - - 22 10 M M M M M M MO MO MO M - MO MO 0 0 0 - -- 45 11 - M - M MO M MO 0 MO MO MO MO MO 0 0 0 0 0 0 35 12 M M M - M M MO MO MO M MO MO MO 0 MO 0 MO MO MO 48 13 M - M M MO M M MO MO MO MO MO MO 0 0 0 MO 0 52 14 M - 0 M M M M M M M 0 MO MO 0 0 0 0 MO 0 0 11 15 - M M M M M M M 0 MO MO MO MO 0 MO 0 0 0 0 19 16 - M M M M M M M M MO MO MO MO MO MO MO 0 0 0 29 - 17 M M M M M MO M M MO MO MO MO MO 0 MO 0 - - - 31 18 M M M M M M M 0 MO MO MO MO MO 0 MO - - - 07 19 - M M M M M M M M M MO MO 0 0 0 0 0 0 0 - so 20 M M M M M M M M M MO 0 0 0 0 0 0 0 0 51 21 MO M M M M M M M M M - MO 0 0 0 0 MO MO 0 - - - 46 22 M M M M M M M MO MO MO 0 0 MO 0 0 0 - 43 23 M M M M M M M M M - MO 0 0 0 0 0 0 0 44 24 M M M MO M M M M M M - MO 0 0 0 0 0 0 0 49 25 M M M M M M M M M M - MO MO 0 MO 0 MO MO 0 38 26 M M M M M M M M M M M MO MO MO 0 0 0 0 0 - 47 27 M M MO M M MO MO M M - MO MO 0 MO 0 0 MO 0 33 28 M M M M M M M MO M M M MO 0 0 0 0 0 0 0 12 29 M M M M M M MO 0 M M - MO MO 0 0 0 0 0 0 - - 10 30 M M M M M M M - - MO MO 0 0 0 0 0 0 40 31 H M M MO M M MO MO M M M MO MO MO MO 0 0 M M 28 32 M M M M M M M M M M M MO 0 MO MO 0 - - - 18 33 M M M M M M M M M M M MO MO MO M MO 0 M 0 - 04 34 M M M M M MO M M M - MO MO MO 0 MO MO MO - 0 39 35 M M M M M M M 0 0 - MO 0 MO MO MO - - OS 36 M M - M M M M M M M MO MO MO MO MO MO - - - 03 37 MO M - M MO M MO M -- - MO MO MO MO MO - - - 27 38 M M M M M M M M 0 MO - MO MO MO MO MO - - - 42 39 M M MO M M M M M M -- MO MO MO MO MO - - - 41 40 M M MO M M M M M M M - MO MO MO MO MO - - - 15 41 M M M M M M M M M M - MO MO MO MO MO - - - 17 42 H M - M M M M M M M - MO MO MO MO MO -- - 16 43 M M M M M M M M M M - MO MO MO MO MO - - - 34 44 M M M M M M M M M M - MO MO MO MO MO - - - 21 45 M M M M M M M M M M M MO MO MO MO MO - - - 20 46 H M M M M M M M M M M MO MO MO MO MO - - - 37 47 M M M M M M M M M M M MO MO MO MO MO - - - 23 48 M M M M M M M M M M M MO MO MO MO MO - -- 26 49 MO M - M M M M M M M M M MO MO 0 MO - - - 25 so M M - M M M M M M M M M MO MO MO - - - - 14 51 M M M M M M M M M M M M MO MO MO MO ------30 52 M M M M M M M M M M M M M M M M 88 5.2 Language choice and respondent characteristics The table in Appendix 8 shows the characteristics of the respondents . Table 5.4 shows the following characteristics arranged in a matrix with the respondents ordered according to their language choice ranking (as determined by the scalogram in Table 5.1): their age ; their sex; which church they go to; where their home is; if & where they work; where they spent their early life; where they spend their free-time; the grade they attained at school . The results in Table 5.4 indicate that a consistent relationship exists between some of the respondent characteristics and the respondent ranking . Tables 5.5 to 5.13 show the relationships between characteristics and language ranking. 89 Table 5.4: Characteristics of respondents (ranked) Resp Language Age Sex Church Home Work Early No. ranking Free Grade life time 06 1 17 F u Saroa 32 0 BB 10 2 30 F u HB 0 0 36 3 12 26 F u Hohola Gordon BB 0 12 09 4 44 F c Kone Kone 13 BB 0 12 5 26 F c Gordons POM 0 0 01 6 13 30 M u Kone POM 0 0 12 02 7 26 M c Wewak Wewak HB 08 BB 8 8 32 F u Toks POM BB BB 10 24 9 35 M u Rabaul HB 22 0 11 10 47 M u HB BB HB 13 45 11 52 M c POM POM BB/0 0 12 35 12 27 F u Saroa Saroa HB HB 8 48 13 38 M u Kaevaga POM HB BB 13 52 14 32 M u HB Kone BB 0 13 11 15 26 F u POM POM BB 0 12 19 16 28 F u Barakau POM HB 0 13 29 17 38 M c HB HB HB 13 31 18 54 M u BB BB HB 5 07 19 30 M u BB Boroko HB BB 10 50 20 25 M u HB POM HB BB 10 51 21 46 M u HB POM HB BB 13 46 22 26 F u BB Gordon BB BB 10 4 3 23 26 F u HB POM BB BB 10 44 24 28 M u BB 9 Mile BB BB 6 49 25 33 M u HB POM HB HB 10 38 26 29 F c HB POM HB HB 10 47 27 23 M L HB POM BB HB 11 33 28 55 M u Boroko Boroko BB 0 13 12 29 40 F u POM POM BB 0 10 10 30 26 M c HB Boroko BB HB 10 40 31 49 M u BB POM BB BB 13 28 32 52 M u BB HB HB 6 18 33 59 F u BB BB BB BB 8 04 34 28 M u HB POM BB BB 12 39 35 26 F u HB HB HB 8 05 36 42 M u BB BB BB 9 03 37 25 M c HB BB HB 10 27 38 35 F u HB HB BB 9 42 39 36 M u BB HB BB 6 41 40 42 M u HB HB HB 10 15 41 23 F u BB BB HB 10 17 42 28 F u HB HB HB 5 16 43 31 M u BB HB BB 13 34 44 23 F u HB BB BB 10 21 45 53 M u BB BB HB 9 20 46 40 F u HB HB HB 10 37 47 73 F u HB HB HB 6 23 48 44 M u BB BB HB 8 26 49 69 M u HB HB HB 3 25 50 52 F u HB HB HB 6 14 51 42 M u HB HB HB 6 30 52 60 F u HB HB HB 6 Key : HB • Hanuabada u = United Church L = Latter Day 0-Outside c = Catholic Church Saints 90 5.2.1 Language rank and age Table 5.5 seems to show that the those under 40 year of age have a lower language rank than those aged 40 and above . The chi-square test indicates that there is not a statistically significant relationship at the .OS level between these age types and these language rank types. Table 5.5: Language ranks by age of respondents Language Age of Respondents Ranks 17-39 40-73 % % Low (1-13) 31.25 15.00 Mid (14-34 ) 43.75 35.00 High (35-52 ) 25.00 50.00 Totals 100.00 100.00 (N) (32) (20) chi-square = 3.756; d.f. = 2; n.s. p > .OS However, calculation of Spearman 's rank order correlation coefficient indicates a fairly weak but nevertheless statistically significant relationship at the .02 level between age rank and language rank (coefficient = .325; d.f. - 51; p < .02). 91 5.2.2 Language rank and gender The results in Table 5.6 show no support for a consistent relationship between gender and language rank type . The largest group (48 per cent ) of males is of the mid-rank language type (i.e. , with rank 14-34 ), whereas the largest group (39 per cent ) of females is of the high-rank language type (i.e. , with rank 3S-S2 ). Tabl e 5.6: Language ranks by gender of respondents Language Gender of Respondents Ranks Male Female % % Low (1-13) 20.69 30.43 Mid (14-34 ) 48.28 30.43 High (3S-S2 ) 31.03 39.13 Totals 100.00 100.00 ( N) (29) (23) chi-square = 1.741; d. f. = 2; n.s. p > .OS The chi-square test indicates that there is not a statistically significant relationship at the .OS level between gender and language rank type (chi-square = 1.741; d.f. • 2; n.s. p > .OS) . However, it is interesting to note that ranks 1-S are occupied by females. 92 5.2.3 Language rank and religion The results in Table 5.7 show the distribution of the language types by religion. As can be seen, 43 of the respondents were affiliated with the Uniting Church and only eight with the Catholic Church (the remaining respondent was affiliated with the Church of the Latter Day Saints). The distribution of language rank types is interesting and shows some support for a consistent relationship between religion and language rank type . However, the three cells in the •catholic" column have an expected frequency of less than five and so it is not possible to obtain a valid chi-square statistic . Table 5.7: Language ranks by religion of respondents Language Religion Ranks United Catholic % % Low (1-13) 20.93 50.00 Mid (14-34 ) 39.53 37.50 High (35-52 ) 39.53 12.50 Totals 100.00 100.00 (N) (43) ( 8) 93 5.2.4 Language rank and home The results in Table 5.8 seem to indicate that people who live in the village are more likely to have a higher language rank and those who live in the city are more likely to have a lower language rank. Less than eight per cent of those with their home in the village were of the low-rank language type (with ranks 1-13), as opposed to 71 per cent of those with their home in the city. However, the two cells in the "City" column have an expected frequency of less than five and so it is not possible to obtain a valid chi-square statistic. Table 5.8: Language ranks by home of respondents Language Location of Home Ranks Village City % % Low (1-13) 7.89 71.43 Mid (14-34 ) 44 .74 28.57 High (35-52 ) 47.37 Totals 100.00 100.00 ( N) (38) (14) 94 5.2.5 Language rank and work The results in Table 5.9 show that people who do not work are more likely to have a high language rank and those who do work are more likely not to have a high language rank. Seventy two per cent of non-workers were of the high-rank language type (with ranks 35-52 ), whereas none (0 per cent ) of the workers were of this type . Table 5.9: Language ranks by work of respondents Language Not Ranks Working Working % % Low (1-13) 33.33 16.00 Mid (14-34 ) 66 .67 12.00 High (35-52 ) 72 .00 Totals 100.00 100.00 (N) (27) (25) chi-square = 30.606; d.f. = 2; p < .001 The chi-square test indicates that there is a statistically significant relationship at the .001 level between work and language rank (chi-square = 30.606; d.f. = 2; p < .001). 95 5.2.6 Language rank, home and work The results in Table 5.10 seem to show that: people who live in Hanuabada and do not work are more likely to have a high language rank; those who live in Hanuabada and work are more likely to have a mid language rank; those who live out of Hanuabada and work are more likely to have a low language rank. However, three of the cells in the table have an expected frequency of less than five and so it is not possible to obtain a valid chi-square statistic. Table 5.10: Language ranks by home & work of respondents Living out Living in Living in Lanquage Hanuabada Banuabada Banuabada & Ranks & Working & Working Not Working % % % Low (1-13) 71.43 6.67 8.70 Mid (14-34 ) 28.57 93.33 13.04 High (35-52 ) 78.26 Totals 100.00 100.00 100.00 ( N) (14) (15) (23) 96 5.2.7 Language rank and early life The results in Table 5.11 seem to indicate that people who were born and bred in the village are more likely to have a mid-to-high language rank and those who were born and bred out of the village are more likely to have a low language rank. Just over 17 per cent of those born and bred in the village were of the low-rank language type (with ranks 1-13), compared to all (100 per cent ) of those born and bred out of the village. However, the three cells in the "Out village" column have an expected frequency of less than five and so it is not possible to obtain a valid chi-square statistic. Table 5.11: Language ranks by early life of respondents Language Born and bred Ranks In village Out village % % Low (1-13) 17.02 100.00 Mid (14-34 ) 44.68 High (35-52 ) 38.30 Totals 100.00 100.00 ( N) (47) ( 5) 97 5.2.8 Language rank and free time The results in Table 5.12 show some support for a consistent relationship between language rank and where people spend their free time. Those who spend half (or more ) of their free time in the village are more likely to have a mid-to-high language rank and those who spend half (or more ) of their free time out of the village are more likely to have a low-to-mid language rank. Fifteen per cent of those spending their free time in the village were of the low-rank language type (with ranks 1-13), compared to 58 per cent of those spending their free time out of the village . However, the three cells in the "Out village" column have an expected frequency of less than five and so it is not possible to obtain a valid chi-square statistic . Table 5.12: Language ranks by free time of respondents Language Spends half free time Ranks In village Out village % % Low (1-13) 15.00 58.33 Mid (14-34 ) 40.00 41.67 High (35-52) 45.00 Totals 100.00 100.00 (N) (40) (12) 98 5.2.9 Language rank and grade Table 5.13 shows that of those who completed grades 1-10, only 12 per cent were of the low-rank language type , compared to 47 per cent of those who completed grades 11-13. On the other hand, 51 per cent of those who completed grades 1-10 were of the high-rank language type, compared to five per cent of those who completed grades 11-13. Table 5.13: Language Ranks by Grade of Respondents Language Grade achieved Ranks 1-10 11-13 % % Low (1-13) 12 .12 47.37 Mid (14-34 ) 36.36 47.37 High (35-52 ) 51.52 5.26 Totals 100.00 100.00 ( N) (33) (19) chi-square = 13.805; d. f. = 2; p < .01 The chi-square test indicates shows a statistically significant relationship at the .01 level between school grade and language rank type (chi-square = 13. 805; d.£ ••2; p < .01). Spearman 's rank order correlation coefficient indicates a statistically significant 99 relationship at the .005 level between grade rank and language rank (coefficient = -.541; d.f. =51; p < .005). 5.3 Network villageness and respondent characteristics Appendix 10 shows the villageness score for each respondent 's network calculated as a percentage of the maximum score . The results, described in the following sections, indicate that consistent relationships exist between some of the respondent 's characteristics and the villageness of the respondent 's network . Because of the small number of respondents with certain characteristics, it is not possible to obtain valid chi-square statistics for: religion, home , early life, and free-time. 5.3.1 Network villageness and age Calculation of Spearman 's rank order correlation coefficient indicates a statistically significant relationship at the .02 level between age rank and network villageness rank (coefficient = .250; d.f. =51; p < . OS ). 5.3.2 Network villageness and gender The results support the hypothesis that there is no consistent relationship between gender and network villageness. The chi-square test indicates that there is 100 not a statistically significant relationship at the .05 level between gender and network villageness type (chi-square = 1.804; d.f. = 2; n.s. p > .05). 5.3.3 Network villageness and home Table 5.14 seems to indicate that people who live in the village are more likely to have a higher network villageness score and those who live in the city are more likely to have a lower network villageness score . Just over 13 per cent of those with their home in the village had a low network villageness score, compared with 79 per cent of those with their home in the city. However, the three cells in the wcityw column have an expected frequency of less than five and so it is not possible to obtain a valid chi-square statistic . Table 5.14: Network villageness by home of respondents Network Location of Home Villageness Village City % % Low (15-42) 13.16 78.57 Mid (43-71) 39 .47 21.43 High (72-97 ) 47.37 Totals 100.00 100.00 (N) (38) (14) 101 5.3.4 Network villageness and work The results in Table 5.15 show a consistent relationship between work and network villageness score . Sixty four per cent of non-workers had a high network villageness score whereas only seven per cent of the workers had a score of this type. The chi-square test gives a statistically significant relationship at the .001 level between work and language rank (chi-square = 19 .090; d.f. = 2; p < .00 1). Table 5.15: Network villageness by work of respondents Network Not Villageness Working Working % % Low (15-42) 48.15 12 .00 Mid (43-71) 44.44 24 .00 High (72-97) 7.41 64.00 Totals 100.00 100.00 ( N) ( 27 ) (25) chi-square = 19.090; d. f. = 2; p < .001 102 5.3.5 Network villageness , home and work The results in Table 5.16 seem to show that: people who live in Hanuabada and do not work are more likely to have a high network villageness rank; those who live in Hanuabada and work are more likely to have a mid network villageness rank; those who live out of Hanuabada and work are more likely to have a low network villageness rank. However, four of the cells in the table have an expected frequency of less than five and so it is not possible to obtain a valid chi-square statistic . Table 5.16: Network villageness by home & work of respondents Living out Living in Living in Network Banuabada Hanuabada Hanuabada & Villageness & Working & Working Not Working % % % Low (15-42) 78.57 20.00 8.70 Mid (43-7 1) 21.43 66.67 21.74 High (72-97 ) 13.33 69.57 Totals 100 .00 100.00 100.00 ( N) (14) (15) (23) 103 5.3.6 Network villageness and grade The results in Table 5.17 show that of those who completed grades 1-10, 12 per cent had a low network villageness score (14-42 ), compared to 63 per cent of those who completed grades 11-13. On the other hand , 48 per cent of those who completed grades 1-10 had a high network villageness score (72-97 ), compared to 11 per cent of those who completed grades 11-13. Table 5.17: Network villageness by grade of respondents Network Grade achieved Villageness 1-10 11-13 % % Low (15-42) 12 .12 63.16 Mid (43-71) 39 .39 26.32 High (72-97 ) 48.48 10.53 Totals 100.00 100.00 ( N) (33) ( 19 ) chi-square = 15.822 ; d. f. = 2; p < .001 The chi-square test indicates that there is a strong and statistically significant relationship at the .001 level between school grade and network villageness score (chi-square = 15.822; d.f. = 2; p < .00 1). 104 Calculation of Spearman 's rank order correlation coefficient indicates a strong and statistically significant relationship at the .005 level between grade rank and network villageness rank (coefficient = -.606; d.f. = 51; p < .005). 5.4 Language choice and villageness Appendix 10 shows the villageness score for each respondent calculated as a percentage of the maximum score . Table 5.18 shows the respondent villageness scores converted to rankings. In this table , the respondents have been ordered according to their language ranking (as determined by the implicational scale in Table 5.1). 105 Table 5.18: Comparison of rankings and types V i 1 1 a g e n e s s L an u Resp g a g e Respondent Network rank type Ho. rank type rank 6 14.5 1 3.0 1.0 1 32 7.5 1 12 .5 2.0 1 36 3.0 1 7.0 3.0 1 9 7.5 1 8.0 4.0 1 13 3.0 1 4.0 5.0 1 1 3.0 1 12 .5 6.0 1 2 11.0 1 12 .5 7.0 1 8 14 .5 1 15.5 8.0 1 24 3.0 1 25.0 9.0 1 22 41.0 2 12 .5 10.0 1 45 11.0 1 18.0 11.0 1 35 23.0 1 21.5 12 .0 1 48 14 .5 1 23.0 13 .0 1 52 14 .5 1 5.5 14.0 2 11 7.5 1 2.0 15.0 2 19 3.0 1 5.5 16 .0 2 29 41.0 2 41.5 17 .0 2 31 41.0 2 19 .0 18.0 2 7 23.0 1 26.5 19 .0 2 50 23.0 1 33 .5 20.0 2 51 23.0 1 15 .5 21.0 2 46 23 .0 1 24.0 22 .0 2 43 23.0 1 21.5 23.0 2 44 23.0 1 35 .0 24.0 2 49 23.0 1 26.5 25.0 2 38 23.0 1 29.5 26.0 2 47 23.0 1 33 .5 27 .0 2 33 11.0 1 1.0 28.0 2 12 7.5 1 10.0 29.0 2 10 23. 0 1 32 .0 30.0 2 40 23 .0 1 9.0 31.0 2 28 41.0 2 52 .0 32 .0 2 18 41.0 2 40.0 33 .0 2 4 23.0 1 20.0 34.0 2 39 41.0 2 17 .0 35 .0 3 5 41.0 2 28.0 36 .0 3 3 41.0 2 36 .5 37 .0 3 27 41.0 2 31.0 38.0 3 42 41.0 2 39 .0 39 .0 3 41 41.0 2 43 .0 40.0 3 15 41.0 2 36 .5 41.0 3 17 41.0 2 38.0 42.0 3 16 41.0 2 41.5 43.0 3 34 41.0 2 45.0 44.0 3 21 41.0 2 29 .5 45.0 3 20 41.0 2 45.0 46.0 3 37 41.0 2 45.0 47 .0 3 23 41.0 2 48.5 48.0 3 26 41.0 2 48.5 49.0 3 25 41.0 2 48.5 50.0 3 14 41.0 2 51.0 51.0 3 30 41.0 2 48.5 52 .0 3 106 Table 5.19 shows the relationship between respondent 's villageness (shown as one of two categories) and language rank (shown as one of three language rank types ). The results show considerable support for a consistent relationship between respondent 's villageness and language rank. Those who have a high villageness score are more likely to have a high language rank and those who have a low villageness score are more likely to have a low-to-mid language rank . Only four per cent of those with a villageness score of 81-100 were of the low-rank language type, compared to 41 per cent of those with a villageness score of 10-80. On the other hand , 78 per cent of those with a villageness score of 81-100 were of the high-rank language type, but none of those with villageness scores of 10-80 were of this type . Table 5.19: Language ranks by villageness of respondents Language Villageness of respondent Ranks 10-80 81-100 % % Low (1-13 ) 41.38 4 . 35 Mid ( 14-34) 58 . 62 17.39 High (35-52 ) 78.26 Totals 100.00 100.00 ( N) (29) (23) chi-square - 35. 131; d. f. = 2; p < .001 107 The chi-square test indicates that there is a strong and statistically significant relationship at the .001 level between respondent 's villageness score and language rank type (chi-square = 35. 131; d.f. = 2; p < .00 1). Calculation of Spearman 's rank order correlation coefficient indicates a strong and statistically significant relationship at the .005 level between respondent 's villageness rank and language rank (coefficient = .797; d.f. =51; p < .005). Table 5.20 shows the relationship between respondent 's network villageness (shown as one of two categories) and language rank (shown as one of three language rank types ). The results show considerable support for a consistent relationship between respondent 's network villageness and language rank. Table 5.20: Language ranks by villageness of network Language Villageness of network Low Mid High Ranks (15-42) (43-71) (71-97 ) % % % Low (1-13 ) 56.25 22.22 Mid (14-34 ) 43.75 55.56 22.22 High (35-52 ) 22.22 77.78 Totals 100.00 100 .00 100.00 {N) (16) (18) (18) chi-square • 29 .371; d. f. = 2; p < .001 108 Those who have a high network villageness score are more likely to have a high language rank and those who have a low network villageness score are more likely to have a mid-to-low language rank. None (0 per cent ) of those with a network villageness score of 71-97 were of the low-rank language type (with ranks 1-13), compared to 56 per cent of those with a network villageness score of 15-42. Conversely, 78 per cent of those with a network villageness score of 71-97 were of the high-rank language type (with ranks 35-52 ), but none (0 per cent ) of those with network villageness scores of 15-42 were of this type . The chi-square test indicates that there is a strong and statistically significant relationship at the .001 level between network villageness score and language rank type (chi-square = 29.371; d.f. = 2; p < .00 1). Calculation of Spearman 's rank order correlation coefficient indicates a strong and statistically significant relationship at the .005 level between respondent 's network villageness rank and respondent 's language rank (coefficient = .780; d.f. = 51; p < .005 ). 5.5 Language rank types Table 5.21 shows each of the respondent and network characterisitics dealt with in the previous sections, and relates these to the three language rank types . 109 Table 5.21 : Characteristics of language rank types LANGUAGE RANK T Y P B S Low Mid High All % % % Aqe Group % 17-39 77 67 44 40-73 62 23 33 56 38 All 100 100 100 100 Ba.e Location In Village 23 81 100 73 Out Village 77 19 27 All 100 100 100 100 Work Working 69 86 52 Not Working 31 14 100 48 All 100 100 100 100 Barly Life In Village 62 100 100 90 Out Village 38 10 All 100 100 100 100 Free Ti.-e In Village 46 76 100 77 Out Village 54 24 23 All 100 100 100 100 School Grade 1-10 31 57 94 63 11-13 69 43 6 37 All 100 100 100 100 Respondent Villaqeness 10-80 92 81 56 81-100 8 19 100 44 All 100 100 100 100 •etvork Villageness 10-59 100 52 6 48 60-100 48 94 52 All 100 100 100 100 Chapter 6 DISCUSSION 6.1 Languages at home The results of Part 1 of this study show that although nearly two thirds of the respondents use only Motu in the home , nearly one third use at least one other language as well as Motu , and the remaining three per cent do not use Motu at all in the home . The fact that so many respondents report that Motu is the language "spoken by everyone in the home and in the village" clearly shows that almost everyone in the community uses it . The degree of usage however, will vary from family to family. Part 1 of the study reported that a few families hardly use Motu in their home at all , for example, respondents 21 and 55 reported using English and another language as the home languages as a result of having parents from different races , and a spouse who was non-Motuan . This seems rather strange because these families live in the village environment and one would expect that the linguistic norms of the Motu-speaking extended family might be imposed on the nuclear family. Here we see examples of the communication function of language choice over-riding the •ocial and cultural functions. 111 For the majority of respondents (approximately 60 per cent) , the reason for using Motu around the home is because it is a sign of their identity . Some of the respondents who did not explicitly mention "identity" as a reason but did say that it was "their language", demonstrated the sense of identity by their intonation and stress on the possessive pronouns in the interviews . For example, Respondent 3 said he used Motu at home "because this is our m:m_ language" and Respondent 5 agreed that "it is our language". Motu gives the speaker a sense of belonging to a particular group. A mother tongue is deeply rooted in the speaker 's personal and cultural identification. The respondents ' views on which languages should be learned at home indicate a language-threatening situation. As can be seen in Figure 6.1, although nearly all (98 per cent ) of the respondents reported using Motu, only 68 per cent thought that it should be learned at home . Only 40 per cent use English at home but 73 per cent indicated it should be learned at home . If more people want to make English the 'home ' language a shift will be inevitable . 112 . 18 ...,., used should be learned 0 � s: 188 I) � s: 0 $).. 88 ., I) r:z: 68 c... 0 ll 48 tn IG � s: ll 28 (.) "" ll ca.. 8 Eng I ish t1otu Language Figure 6.1: Languages used or to be learned at home Just over half of the respondents reported that English should be learned at horne because it will help at school and work {although only five per cent reported this as the reason for it actually being used in the horne currently ). Nine per cent of the respondents said that English should be learned at home because it is a "universal/international language" . 113 6.2 Languages at school All the respondents for Part 1 of the study had attended or were attending school and 42 per cent reported that they had reached grade 10 or over. This figure is more than double the national average, which is 18 per cent of the population age group (National Statistics Office, 1985). This is not surprising because Banuabada is located in the midst of where modernisation first took place and continues to occur. An interesting point was that all the people over 50 years old had only completed grades three to six. Nearly all the respondents (about 95 per cent) acquired Bnglish literacy and used English at scho ol, the majority (82 per cent ) reporting that this was because it is the language of instruction. As regards views about which language should be learned at school , Figure 6.2 shows that a higher percentage (85 per cent ) of respondents favour learning English, rather than Motu , at school . People in Banuabada believe English to be the only language for school "because it is the standard educational language" (Respondent 7). It is noticeable that whereas 40 per cent report using Motu at school, only ten per cent favour learning it at school . Of course, since most children would learn to speak Motu at home , it would be strange to suggest that they would need to learn to speak it at scho ol. But learning a language at school implies learning to write it, and this would not be learned at home . 114 . Ia used should be learned 's: 188 ., ., s: 0 88 Po • ., a: 68 ... 0 ., tD 48 's: ., u 28 Lo ., 0.. 8 l:ne l lah ltotu Language Figure 6.2: Languages used or to be learned at school The more favourable attitude for English than for the local language , expressed in both the horne and school situations, reveals how important English language is for people �n Banuabada. This attitude is again revealed in the respondents' responses to the questions based on the Tok Ples policy of the Papua New Guinean Department of Education . Over 70 per cent of the respondents oppose the use of Motu in pre-school and Grade 1. 115 6.3 Languages at work Eighty seven per cent of the respondents had some member or members of the extended family working in the formal sector earning an income . Families in Hanuabada depend on having someone working in the formal sector as their economy has changed to one that is largely based on cash. Forty-eight per cent use Motu at work: 43 per cent use Motu in combination with other languages and five per cent use Motu only. This last figure probably indicates that the jobs involved are casual . For example, I know a lot of Banuabadans who work in the printery of the Post Courier newspaper on night shifts where they assemble the paper as soon as it comes off the mac hines ready for sale the next morning. There is no written work required and any person in authority may well be a Banuabadan himself or herself , therefore Motu becomes the obvious choice of language . It could also be the case that some of the respondents used •Motu• to mean "Biri Motu ". Despite that, for Banuabada people English is the language of the formal sector of employment . Ninety-seven per cent use English at work; five per cent even aquired literacy in English whilst at work . The fact that nearly all the respondents reported using English at work is not surprising as English is the official language of employment in PNG . 116 6.4 Languages in church Almost every person in Hanuabada is a Christian and religion is taken seriously by most people in the village. Nearly everyone attends church on Sunday every week. Church is therefore an important place or situation where language choice can easily be observed . Apart from the Sunday services, there are plenty of other church activities organised throughout the whole week. Bible Studies , for example, is organised for the youth on one night and for older people on another night. It seems to me that Motu is well and truly alive in churches of the United Church denomination. As indicated by a couple of respondents, Motu is the church language in the United Church (to which 73 per cent of the participants belong), and for some this makes understanding Motu very important . The United church services are conducted predominantly in Motu : the readings , the celebrant 's sermon or homily, the hymns and all prayers are led in Motu . I personally think that the United Church in Hanuabada is perhaps the only social situation that restricts language use to Motu only . On the other hand, the Catholic services are conducted predominantly in English. Motu is used for announcements regarding the community, usually just before the final prayer. Sometimes, but rarely, Motu hymns are sung. There are coDDDunity groups, which, with the help of the nuns (Missionaries of Charity ), voluntarily organise other church activities . It may be the case that these group-oriented 117 activities (e.g., the Legion of Mary ) utilise Motu more than English, but I am not certain as I do not have any relevant data. Apart from the United and Catholic groups , other denominations are present, namely Jehovah 's Witness and the Church of Latter Day Saints. Again I do not know about language use in those churches. However, they are only very small groups . The missions have had a powerful influence in PNG. Historically, in their attempt to convert people quickly and easily, adoption of the local language became necessary. And so the process of firstly educating and ultimately converting people often took place in the vernacular. As a result, the United Church in Hanuabada has become strong and dominant in Motu. Motu (presumably Hiri Motu ) was also chosen as the church lingua franca around some parts of the surrounding region. 118 6.5 The language ecology of Banuabada Language shift studies look not at individual languages but entire bilingual or multilingual situations , seeing the functions of each language as fitting together to make a whole. (McConvell, 1991a, p.1) In PNG , the multitudes of languages exist to serve specific functions in their respective contexts of use. On a broad level, Motu serves the specific function of uniting all the Motuans of the Port Moresby region. Command of the language , which can be either passive or active , identifies the speaker as a Hotuan person , giving him or her an identity. Quite obviously Motu is not the only important language in the community of Banuabada, but it is the one viewed as of great importance by most (83 per cent ) of the respondents. Most of the reasons given for the importance of Motu indicate that the language serves the social function of expressing group identity and bonding. One quarter of the respondents consider Motu to be very important for cultural maintenance whilst just over half (56 per cent ) consider it to be important because it is their mother tongue. Bnglieh is highly regarded (75 per cent said that it is just a• important as Motu ) because it is an international/ universal language . Considered to be the language needed to 119 get good jobs in the formal sector , it is also a lingua franca both within and outside the village, city and national boundaries. Tok Pisin is considered important (by 13 per cent of respondents) because it �s a national lingua franca and makes communication easier with non-Motuans in Port Moresby . None of the respondents indicated that Tok Pisin served the social function of expressing group identity and bonding. It is interesting to compare these attitudes to Tok Pisin with Schmidt 's observations about creoles in Australia. She says that although the development of creoles was triggered by the need of a lingua franca to serve as a commonly shared language, creole languages perform a vital identity function for their speakers by marking their ethnicity/Aboriginality as distinct from English, the language of white society. {Schmidt, 1990, p.112) As far as I know Tok Pisin does not exist as a first language for any of the children in the village of Banuabada . Of course, Tok Pisin is a first language for aany children on the New Guinea side of Papua New Guinea. In Lae, for example, Tok Pisin is fast becoming a creole (Siegel, 1992; Smith, 1992 ). Biri Motu, the local (Papuan) lingua franca was not epecifically mentioned by any of the respondents. Perhaps the respondents take it for granted . Indeed, it is hard to know if in some cases respondents answered "Motu" when they 120 might have meant "Hiri Motu". However, Muhlhausler (1992) suggests that increased support for pidgins based on indigenous languages such as Biri Motu may assist in language maintenance by keeping indigenous and non-traditional contact languages apart . Figure 6.3 shows a comparison of languages used in four different situations ; namely, home , school , work and city . Clearly, Motu is the language of the home and the village . However, it is not used in isolation in these situations; rather, it is one of the many languages utilised. Thirty-six per cent of the respondents reported using two languages in the home situation. Some respondents indicated that Tok Pisin or even English was their home language . English is dominant in other locations , but again, it is not used in isolation. It was reported to be used (although not exclusively) by 95 per cent at school . In fact, 40 per cent reported using Motu in addition to English. At work, English was reported to be used by nearly all of the respondents in paid employment, with 27 per cent using English alone and the rest in combination with other languages. Thirty-nine per cent reported using two languages and 29 per cent reported using three or more languages at work. In the city, 31 per cent said they used three or more languages . 121 . Ho..e School Work Cit� DoMa in Figure 6.4: Language used at home, school, work and city Key: holM! (N•SS) ; school (NeSS) ; work (N�44); city (N;SO) Prom Figure 6.4, one can see the domination of the English language in three out of the four contexts. The home situation is the only one where Motu is used by the majority of respondents. Tok Pisin on the other hand is not as popular as English, but it is used by nearly 50 per cent of 122 respondents both at work and in the city. This shows that Tok Pisin may well be on the way to assuming the roles of being a ready means of communication and a marker of group identification for some of the Banuabadans who work in the city. Because of its institutional, instrumental and formal identification as an official language suitable for wider communication, English is the preferred language in three of the situations. It appears to be the language one needs to acquire in order to be a part of the fast-changing community of Port Moresby. Nevertheless, it is not used in isolation in the various situations . It is used mainly in conjunction with Motu and Tok Pisin (and with other languages that speakers understand ). These figures demonstrate the considerable degree of multilingualism which exists in each of the situations. But it is not just across each situation that many languages are used. Within each situation different languages are often used by the same individual to interact with different interlocutor types; and sometimes the same individual might use several languages to interact with the one interlocutor type (e.g. , 60 per cent of the respondents reported using the three languages Motu , English and Tok Pisin, with the interlocutor type 'friends '). 123 It has been noted that the variety of Motu spoken by younger Hanua.badans is not the same as that spoken by older people in the village. Intrasentential and conversational code switching are more commonly noted in the speech of younger people. The following intrasentential English/Motu code-switching was recorded on my second visit: Speaker: "Second time be reana 1994." Translation: (Second time will be perhaps 1994 .) The following recorded conversation contains examples of both intrasentential and conversational code-switching: M1 : "Berea ." (Very nice . ) M2: "Hmm••. dia oi dibamu aaa ••• safe ani ?" (Umm •• It's not, you know •.erm •• safe isn't it? ) M1: "Very safe ." M2: "Mmm dabanini." (Umm no wonder .) M1: "Very, very safe ." One respondent noted that a lot of English words are heard in the childrens ' speech around the village. Some are borrowed words which have been adapted, e.g., sikuli ("school") and motuka ("motor car" ) and are thus examples of lexical nativisation. Words like tanikiu ("thank you" ) and pilisi ("please") would not have been used in pre-colonial days, but are quite common now. 124 6.6 Speakers • language choices with interlocutors Appendix 9 shows the languages which each respondent reportedly uses with various interlocutor types. The majority (81 per cent) of the respondents reported using at least three languages (mostly Motu , English and Tok Pisin) for this group of interlocutors. Only one respondent reported only using Motu with each of the interlocutor types . Some interlocutor types attracted the use of more languages than others; for example, sixty per cent of the respondents reported using three languages with the interlocutor type 'friends '. The scalogram in Table 5.1 reveals a clear pattern of language choice: from low use of Motu and high use of other languages to high use of Motu and low use of other languages. Knowledge of the identity of the speaker and of the interlocutor-type enables the language choice to be predicted in most cases. The ordering of the interlocutor-types appears to form a scale from maximum to minimum "villageness" . This can be seen by considering this scaled list taken from Table 5.1: grandparents > village seller > traditional doctor > trade store assistant > parents > siblings > spouse > children > grandchildren > bank assistant > government of ficial > co-worker > supervisor. 125 As was seen, six respondents had four or more 'errors ' (language choices which did not fit within the scaled matrix ). Examination of the characteristics of these respondents to some extent explains their "unscalability" : Respondent 1 is a male who was brought up outside of Banuabada, is married to a non-Motuan, now lives and spends free time outside of Hanuabada , but attends ceremonies in Hanauabada; Respondent 32 is a female who was brought up outside of Hanuabada , has a non-Motuan parent (who is a professional person), is married to a non-Motuan, is unemployed, now lives and spends free time outside of Hanuabada and does not attend ceremonies . Given this profile, it is difficult to see why this respondent should choose Motu/Other (rather than only Other ) for the interlocutor types: 'town shop assistant ', 'bank assistant ', and 'government official '; Respondent 35 is a female who was brought up in Banuabada and is married to a Hanuabadan (which explains why she chose only Motu for spouse). Her husband works in the United Church of a non-Motuan village and she does supportive work in that church (clearly she must be working with other Motuans in the church since Motu/Other was selected for the work situation). She lives in the village where she and her husband work, but spends free time and attends ceremonies in Hanuabada; 126 Respondent 40 is a male who was brought up in Hanuabada and who now works in the United Church in Waigani (a suburb of Port Moresby) , but spends his free time and attends ceremonies in Hanuabada . Although he reports Hanuabada as his home , a house is provided for him as the Church Minister 1n Waigani. Thus, his reported usage of Motu/Other for 'neighbours' and 'trade store assistant ' could be because he is referring to the Waigani setting; Respondent 47 is a male outsider who is married to a woman from Hanuabada and he lives in his wife 's clan. He and wife belong to the Church of Latter Day Saints (which may explain why Motu/Other was chosen for 'fellow worshipper '). He is employed in the city and therefore works out of the village. He spends his free time and attends ceremonies in Hanuabada; Respondent 51 is a male whose mother is Motuan and father is European (which explains why Motu/Other was chosen for 'grandparents ' and perhaps why only Other was selected for 'town shop assistant' ). He is married to a woman from Elevala - the other section of the Poeporena village ). He lives in Hanuabada, works in one of the city suburbs, and spends his free time and attends ceremonies in Hanuabada . Fifty per cent of his network members are co-workers. 127 The speaker variables which correlate with and perhaps determine the ordering of the scalable respondents in the scalogram will be dealt with in the following subsections . 6.7 Language choice and respondent characteristics Spearman 's rank order correlation coefficient indicates a statistically significant relationship between age rank and language rank - the older the person, the higher the language rank. There is also a significant relationship between work and language rank. The results in Table 5.9 show that people who do not work are more likely to have a high language rank and those who do work are more likely not to have a low-to-mid language rank. There is a significant negative correlation between grade and language rank. The results in Table 5.12 show that those who did not go beyond Grade 10 at school are more likely to have a mid-to-high language rank and those who have completed higher grades are more likely to have a low-to-mid language rank. There seems to be some support for consistent relationships between home , early life, free time and language rank, but owing to the small numbers involved, it was not possible to obtain a valid chi-square statistic. 128 However, the results do seem to show that people who live in the village are more likely to have a higher language rank than those who live in an urban environment . Also, people who were born and bred in the village are more likely to have a higher language rank than those who were born and bred out of the village. Similarly, the results seem to show those who spend half (or more ) of their free time in the village are more likely to have a higher language rank than those who spend half (or more ) of their free time out of the village. 6.8 Lanquage choice and villageness There is a significant relationship between respondent 's villageness and language rank: the higher the villageness score, the higher the language rank. Similarly, there 1s also significant correlation between the respondent 's network 's villageness and language rank: the higher the network villageness, the higher the language rank. The Spearman's rank order correlation coefficients indicate that the respondent 's villageness is marginally better as'a predictor than the network villageness. This is contrary to Gal 's findings (1979, pages 183-4 ). 129 6.9 Language rank types and life modes There would seem to be three broad 'types ' of language users: 'low-rank language type' , 'mid-rank language type' , and 'high-rank language type '. Each language type has an associated set of characteristics which describe those individuals which belong to that type . The types and their characteristics can be described as follows : 6. 9. 1 Low-rank language type The low-rank language type comprises respondents who use Motu on its own with about one third of the interlocutors, use mixed Motu/other with another third, and use at least one other language {and no Motu ) with the remaining third. People who fall into this language-type are more likely to be under 40 years of age, to live out of the village, to be working , and to have achieved a school grade higher than grade 10. Just over half will spend their free time out of the village. Although nearly two thirds of the respondents in this type spent their early life in the village, all the respondents who spent their early life out of the village were of this low-rank language type . The villageness of people in this type will be low, as will the villageness of their networks . 130 6.9.2 Mid-rank language type The mid-rank language type comprises respondents who use Motu with just over half of the interlocutors , use Other with nearly one third, and mixed Motu/Other with the remainder. All the people who fall into this language-type will have spent their early life in the village. They are more likely to be working , but they are also more likely to live in the village and spend their free time in the village . Two thirds of them will be under 40 years of age. Just over half of the people in this type will have achieved a school grade higher than grade 10. The villageness of people in this type will be low, but nearly half will have networks with high villageness. 6.9.3 High-rank language type The high-rank language type comprises respondents who do not interact with interlocutors at work, because none of them are engaged in paid employment . They use Motu on its own with just over two thirds and mixed Motu/Other with just under one third of the interlocutors with whom they interact . All the people who fall into this language-type will have spent their early life in the village. All of them will live in the village and spend their free time in the village. Just over half will be over 40 years of age. Hardly any of the people in this type will have achieved a 131 school grade higher than grade 10. The villageness of all the people in this type will be high and nearly all will have networks with high villageness . 6.9.4 Life-modes The characteristics associated with each of the language types outlined above suggest that it might be useful to divide the Hanuabada community into three different life-modes. H�jrup (1983) based his account of life-modes on his Marxist analysis of the social and economic structures of western European countries. Thus, for H¢jrup (1983, p.47), the modes of production and consumption are the "fundamental societal structures which split the population into fundamentally different life-modes". The division of any population into these subgroups is a "large-scale and ultimately economically driven process" (Milroy and Milroy, 1992, p.18). Of course, Hanuabadan society is very different to that found in western Europe , and so one should not necessarily expect the same concepts and 'analyses to be useful. For Banuabada , it would seem useful to adopt the concept of 'life-modes ', but not necessarily adopt the explanation of their origin given by H¢jrup (1983). The appropriate life-modes for Hanuabadans seem in part economically determined (as in H�jrup, 1983) but also in part determined by location of home . Thus, in Hanuabada, 'life-mode L' compriaea those who live (and most probably work) out of the 132 village. 'Life-mode M' comprises those who live in the village, but work out of the village . 'Life mode-H ' comprises those who live in the village and do not work . Bach life-mode in Hanuabada gives rise to a particular set of social, cultural and linguistic behaviours, and also gives rise to a particular network structure which enforces these social , cultural and linguistic norms . Life-mode H has the characteristics which are described for the high-rank language type . In particular, people of this life-mode have social networks with high villageness. In many respects this life-mode is similar to Hojrup 's life-mode 1 in that it exhibits a "close-knit type of network structure and a solidarity ethic" (Milroy and Milroy, 1992, p.20). Life-mode M has the characteristics which are described for the mid-rank language type. In particular, people of this life-mode will have low villageness, but some will have networks with high villageness and some will have networks with low villageness . In some respects this is similar to a;jrup's life-mode 2. But Milroy and Milroy (1992, p.20) note that the variability in the extent to which workers in a;jrup 's life-mode 2 have close-knit network ties and a solidarity ethic is a function of the wage level . Life-mode L has the social, cultural, linguistic and network characteristics which are described for the low-rank language type . In particular, people of life-mode L will have networks with low villageness. Like the people of 133 B�jrup's life-mode 3, those in life-mode L "will be socially and geographically mobile as they pursue their careers, forming many loose ties, particularly of a professional kind, through which innovations and influence may be transmitted" (Milroy and Milroy, 1992, p.21) . Thus, the three life-modes proposed for Banuabada are similar to those of B�jrup, and it may well be that further research would reveal a clear relationship between them. Chapter 7 CONCLUSION Banuabada is a multilingual community for which Motu is the local or indigenous language. "Overlapping speech communities" (Saville-Troike , 1982) are common in Banuabada. e.g., almost all respondents reported using Motu in the home , and English at school and/or work . The choice of one code rather than the other is obviously related to the situation (Wardhaugh, 1986) and indicates that "people alter their norms for speech behaviour to conform to the appropriate speech community, by adding , substracting, and substituting rules of communicative behaviour" (Fasold, 1990, p.42). Most families in Hanuabada are dependent on having someone working in the formal sector and earning an income as their economy is largely a cash economy . Acquiring the English language is regarded as an advantage and associated strongly with getting well paid employment . The fact that English 1s the official language of business and administration encourages this attitude. The reasons for learning and using English given by the respondents in this study can all be classed as 'instrumental' and be seen to relate to social mobility and 135 improvement in employment prospects in the formal sector. They are associated with status and prestige, and to this extent we can conclude that English is regarded as a prestige language by a lot of people from Hanuabada . Renck (1990) comments on this as being a cargo cult attitude towards the learning of English, i.e., the attitude that it will bring power and wealth. But like so many other cargo cults, the learning of English and the education system generally have failed to deliver the goods to most Papua New Guineans (Swatridge , 1985 ). It has been maintained by some educationalists that teaching in the vernacular is the best medium for education (UNESCO, 1953 ). It could be argued that if vernacular or bilingual education is adopted in Hanuabada, the maintenance of Motu will be guaranteed , and local participation and autonomy acquired. Also, children will perform better academically at school if allowed to mature in their mother tongue (Cummins, 1978a & 1978b). This has already been shown with the Viles Tok Ples schools in Bougainville (Ahai, 1984 ; Saovana-Spriggs, 1992). There have been several attempts to develop more Tok Ples education programs in Papua New Guinea (Malone & Malone, 1987; Stringer & Faraclas , 1987 ) and, as already noted, the Papua New Guinea Education policy now supports such programs . However, there is little support in Hanuabada for the Papua New Guinea Education policy on the introduction of Tok Ples programs into schools; most of the respondents in Banuabada favour learning English, whereas only a few favour 136 learning Motu, at school. In order to ensure that the indigenous or local language is not replaced by a foreign language, a language maintenance policy is required which has not only political and educational support , but also has broad community support . Although 97 per cent reported 'using ' Motu at home , only 67 per cent raised the issue of 'learning ' it at home , and whereas only 40 per cent 'use ' English at home , 72 per cent indicated it should be 'learned' at home . This desire for English to be learned at home together with the desire for it to be the medium of instruction throughout school, creates a language-threatening situation. If more people want to make English the home language and very few want Motu to be learned at school, then a shift will be inevitable . As we have seen, English is regarded as prestigious because it is the language of the school , employment and to an extent the language used in the city as well as the language for international communications. However, there were some positive attitudes expressed for Motu ; indeed, Motu is viewed as of 'great importance ' by most (83 per cent ) of the respondents. Over two thirds of the respondents in the sample gave reasons for using Motu that implied a sense of both personal and cultural identity - "this is our own language". 137 The study demonstrates that a considerable degree of multilingualism exists in each of the situations. But it is not just across each situation that many languages are used . Within each situation different languages are often used by the same individual to interact with different interlocutor types, e.g. , in the home with siblings, spouse, grandparents, parents and children. The majority of the respondents reported using at least three languages (mostly Motu, English and Tok Pisin) for the interlocutor-types chosen for the study. There would seem to be three broad 'types ' of language users : 'low-rank language type', 'mid-rank language type ', and 'high-rank language type '. The low-rank language type comprises respondents who use Motu on its own the least . The mid-rank language type comprises respondents who use Motu on its own slightly more . The high-rank language type comprises respondents who use Motu on its own more than the other types, and do not use any other language alone. Associated with each language type is a 'life-mode ', which determines the social , cultural and linguistic behaviour, and also the networks of the individuals of that mode. The life-modes are basically determined by where the person lives and whether or not the person works. Thus, 'life-mode L' comprises those who live (and most probably work) out of the village, and who belong to the low-rank language type. 'Life-mode M' comprises those who live in the village, but work out of the village, and who belong to the mid-rank 138 language type. 'Life mode-H ' comprises those who live in the village and do not work , and who belong to the high-rank language type . Further research is required using a larger sample (to enable statistical analysis) so as to establish the relationships between life-modes, social and cultural differences, network structures , and language choices . Some individuals have social networks comprising a 'village cluster ' and a 'work cluster' - such individuals may well serve as the 'weak ties ' through which linguistic innovation is introduced . Further research is required on this. It will also be interesting to administer the same study at five year intervals to plot the changing sociocultural and linguistic situation of the multilingual speech community of Banuabada . 139 BIBLIOGRAPHY ABAI , N. (1984 ) "A case for vernacular education in Papua New Guinea". Yagl -Ambu , 11: 25-38 . AIKIO, M. (1992) "Are women innovators in the shift to a second language : a case study of Reindeer Sami women and men". International Journal of the Sociology of Language, 94: 43-61. ALATIS, J.E. (ed. ) (1978) International Dimensions of Bilingual Education . Washington : Georgetwon University Press. AMERY, R. (1985) A New Diglossia : Contemporary Sp eech Va rieties at Yirrkala, N.E. Arnhem Land. M.A. Thesis, Australian National University . BALDAUF, R. & LUKE , A. (eds.) (1990) Language Planning and Education in Australasia and the South Pacif ic. Clevedon, England: Multilingual Matters . BARNES , J.A. (1954 ) "Class and committees in a Norwegian island parish". Human Relations, 7: 39-58. BBLSBAW, c.s. (1957 ) The Great Village . London : Routledge & Kegan Paul. BLOM, J. & GUMPERZ , J.J. (1972 ) "Social meaning in linguistic structure: code switching in Norway ". In Gumperz & Hymes (eds. ), 1972: 407-434 . BORTONI-RICARDO, S.M. (1985) The Urbanisation of Rural Dialect Sp eakers: A Sociolin�istic Study in Brazil. Cambridge : Cambridge Univers1ty Press. BOTT, E. (1955) "Urban families : conjugal roles and social networks". Human Relations, 8: 345-384. BOTT, E. (1957 ) Family and Social Networks : Roles, Norms, and External Relationships in Ordinary Urban Families. London : Tavistock. 140 CLYNE , M � ( 1982 ) Multilingual Australia : Resources, Needs, Pol�c�es. Burnley: Brown, Prior & Anderson. CLYNE , M. (ed. ) (1985) Australia, Meeting Place of Languages, Canberra: Pacific Linguistics. COLLIER, K.J. (1977) "Predicting intelli9ibility: a suggested technique". In Loving & Slmons ( ed. ) , 1977: 253-61. COMBS , M. (1977) "Cultural considerations in language change communication". In Loving & Simons ( ed . ), 1977: 217-32. CRAWFORD, A.L. & LABUI, J. (1980) The Motu People of PNG . PNG: National Cultural Council . CROWLEY, T. & LYNCH, J. (1983) Preliminary Report on Oceanic Languages . UNESCO. CUMMINS , J. (1978a ) "Educational implications of mother tongue maintenance in minority-language groups". The Canadian Modern Language Review, 34: 395-416. CUMMINS , J. (1978b) "Bilingualism and the development of metalinguistic awareness". Journal of Cross-Cultural Psychology, 9: 131-49. DEVLIN , B. (1986a) Language Maintenance in a Northeast Arnhem Land Settlement . Ed.D. thesis, Columbia University. DEVLIN , B. (1986b) "Some issues relating to vernacular language maintenance ". Talk given to the Northern Territory Institute for Educational Research. June 19. Darwin: Northern Territory Institute for Educational Research. DIL, A. (ed. ) (1974 ) The Ecology of Languages: Essays by Einar Haugen . Stanford: Stanford University Press . DIXON, R.M.W. (1991) "The endangered languages of Australia, Indonesia and Oceania". In Robins & Uhlenbeck (eds.), 1991. 141 DRESSLER, W. & WODAKLEODOLTER, R. (1977) "Language preservation and language death in Brittany". International Journal of the Sociology of Language, 12: 33-44. DUTTON , T.E. (1976a ) "Austronesian languages : eastern part and south-eastern mainland Papua". In Wurm (ed. ), 1976. DUTTON , T.E. (1976b ) Language and National Development - Long Wanem Rot ? Inaugural Lecture . Waigani: University of Papua New Guinea. DUTTON , T.E. (ed. ) (1992) Culture Change, Language Change: Case Studies from Melanesia. Canberra: ANU Pacific Linguistics Series, C-120. DUTTON , T.E. & MUHLHAUSLER, P. (1991) "Are our languages dying?" Occasional Papers in Language and Literature 1(1): 50-72. Waigani : University of Papua New Guinea . DUTTON, T.E., ROSS, M. & TRYON , D. (eds.) (1992 ) The Language Game : Papers in Memory of Donald c. Laycock. Canberra: ANU Pacific Linguistics Series, C-110. FASOLD, R. (1990) Sociolinguistics of Language. Oxford: Basil Blackwell. FERGUSON, C.A. (1959 ) "Diglossia". Word , 15: 325-340. FBSL, B. (1987 ) "Language death among Australian languages". Australian Review of Applied Linguistics, 10: 12-22. FISHMAN, J.A. (1964 ) "LanguaC}e maintenance and language shift as a field of inqu�ry" . Linguistics, 9: 32-70. FISHMAN, J.A. (1965) "Who speaks what language to whom and when? ". Linguistique, 2: 67-88. FISHMAN, J.A. (1967) " ili�gualism �ith and w thout � . � . .. diglossia; digloss�a w�th and w�thout b�lLngual�sm • Journal of Sociological Issues, 23: 29-38. FISHMAN, J.A. (ed. ) (1968 ) Readings in the Sociology of Language. The Hague: Mouton . 142 FISHMAN , J.A. ( 1970) Sociolinguistics: A Brief Introduct�on . Rowley, Mass: Newbury Bouse. FISHMAN , J.A. (ed. ) (1971) Advances in the Sociology of Language . Volume 1. The Hague: Mouton . FISHMAN , J.A. ( 1972a) The Sociology of Language: An �nterd�sc�pl inary Social Science Approach to Language �n Soc�ety. Rowley, Mass: Newbury House. FISHMAN, J.A. (1972b) "Domains and the relationship between micro- and macrosociolinguistics" . In Gumperz & Hymes (eds. ), 1972: 435-453 . FISHMAN , J.A. (ed. ) (1972c ) Advances in the Sociology of Language. Volume 2. The Hague : Mouton. FISHMAN , J.A. (ed. ) ( 1978) Advances in the Study of Societal Multil�ngualism. The Hague : Mouton . / FISHMAN , J.A. (1989 ) Language and Ethnicity in Minority Sociolinguistic Perspective . Clevedon, Philadelphia : Multilingual Matters . FISHMAN , J.A. (1991) Reversing Language Shift : Theoretical / and Empirical Foundations of Assistance to Threatened Languages . Clevedon, Philadelphia: Multilingual Matters. FISHMAN , J.A. , FERGUSON , C.A. & DAS GUPTA, J. (eds. ) (1968) Langua�e Problems of Developing Nations . New York : John W1ley and Sons . FRANKLIN, K.J. & STEFANIW, R. (1992 ) "The pandanus languages of the Southern Highlands Province, Papua New Guinea". In Dutton et al (ed).), 1992 : 1-6 . GAL, s. (1979) Language Shift . Social Determinants of Linguistic Change in Bilingual Austria. New York. Academic Press Inc . GILES, B. & St CLAIR, R.N. (eds.) (1979) Language and Social Psychology. Oxford: Blackwell. 143 GLADWIN , T. & STURTEVANT, W. (eds . � (1962 ) Anthropology and Huma� Behaviour. Anthropolog�cal Society of Wash�ngton . GRANOVETTER, M. (1973) "The strength of weak ties". American Journal of Sociology, 78: 1360-80. GREENFIELD, L. ( 1972) ''Situational measures of normative language views in relation to person , place and topic among Puerto Rican bilinguals". In Fishman (ed. ), 1972c : 17-35. GUMPBRZ , J. & HYMES, D. (eds. ) (1972) Directions in Sociolinguistics: The Ethnography of Communication . New York: Bolt, Rinehart and Winston. GUTTMAN , L. (1944 ) "A basis for scaling qualitative data". American Sociological Review, 9: 139-50. BALL , R.A. (1955 ) Hands Off Pidgin Engl ish . Sydney : Pacific Publications. HALLIDAY, M. (1978) Language as Social Semiotic. London : Edward Arnold. HAUGEN , E. (1974) "The ecology of language". In Oil (ed. ), 1974: 325-339. BELLER, M. (1988 ) Codeswitching: Anthropological and Linguistic Perspectives. Berlin: Mouton de Gruyter. B�JRUP , T. (1983) "The concept of life-mode : a form-specifying mode of analysis applied to contemporary western Europe". Ethnologia Scandinavica 1-50. HUDSON , R.A. (1980) Sociolinguistics. Cambridge : Cambridge University Press. HYMES , D. (1962 ) "The ethnography of speaking". In Gladwin / 'Sturtevant (eds.), 1962. HYMES, D. (1972) "Models of the interaction of language and social life". In Gumperz & Hymes (eds.), 1972: 35-71. 144 JERNUDD, B. (1968) "Linguistic integration and national development: a case study of the Jebel Marra area ' Sudan". In Fishman et al . (eds.), 1968, 107-118. JESPERSON, 0. (1922 ) Language: Its Na ture, Development and Origin. New York: W.W. Norton . JOHNSON , R.K. (1974 ) "Language policy in PNG" . Research Report 28. Port Moresby : UPNG, Faculty of Education . JOHNSON , R.K. (1977 ) "English in Papua New Guinea". In Wurm (ed. ), 1977: 129-55. KALE, J. (1990) "Language planning and the language of education in Papua New Guinea" . In Bauldauf & Luke ( eds . ) , 199 0 : 182- 196 • KERBMA, P. (1989 ) "Bilingualism in community school education". In Thirwall & Hughes (eds. ), 1989: 3-13. �ULIC� , D. (1992a) Language Shift and Cultural - Reproduction : Socialisation, Self, and Syn cretism in a PaRua New Guinean Village. Cambridge : Cambridge Un�versity Press . KULICK, D. (1992b) "Language shift as cultural reproduction" • In Dutton et al (ed) .), 1992: 7-26. KULICK , D. & STROUD, C. (1992 ) "The structure of the Taiap (Gapun ) language". In Dutton et al . (eds. ), 1992 : 203-226. LABOV, w. (1963) "The social motivation of a sound change". �ord, 19: 273-309. LABOV, w. (1972a) Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. LABOV, w. (1972b ) "On the mechanism of linguistic change". In Gumperz & Hymes (eds. ), 1972 : 512-537. LABOV, W. , COHEN, P., ROBINS , C. & LEWIS, J. (1968) A Study of the Non-standard English of Negro and Puerto Rican speakers in New York City. New York: Columbia University. 145 LAOSA, L. (1975) "Bilingualism in three United States Hispanic groups : contextual use of language by children and adults in their families". Journal of Educational Psychology, 67: 617-27. LAYCOCK, D. (1979 ) "Multilingualism: linguistic boundaries and unsolved problems in Papua New Guinea". In Wurrn ( ed. ) , 19 7 9 • LePAGE , R.B. & TABOURET-KELLER, A. (1985) Acts of Identity: Creole-based Approaches to Language and Ethnicity. Cambridge : Cambridge University Press. LINDGREN , A.R. (1984) "What can we do when a language is dying?" Journal of Multilingual and Multicultural Development, 5: 3-4 . LITHGOW, D. (1992 ) "Language change on Fergusson and Normanby Islands, Milne Bay Province, Papua New Guinea". In Dutton et al (ed).), 1992: 27-47. LITTERAL, R. (1975) "A proposal for the use of pidgin in Papua New Guinea 's education system" . In McElhanon ( ed • ) , 19 7 5 : 15 5-6 5 • LOVING, R. & SIMONS , G.F. (eds.) (1977) Language variation and survey techniques . Workpapers in Papua New Guinea Languages 21. Ukarurnpa, Papua New Guinea: Summer Institute of Linguistics. LUDI , G. (1992 ) "Internal migrants in a multilingual country" . Hultilingua , 11: 45-73. MACKEY, W.F. (1968 ) "The description of bilingualism" . In Fishman (ed. ), 1968: 554-84. MACKEY, W.F. (1978) "The importance of bilingual education models". In Alatis (ed. ), 1978. MALONE , S. & MALONE, D. (1987) Developing Tok Ples Education Programs in PNG. Ukarurnpa , Papua New Guinea : Summer Institute of Linguistics. MATTHEWS D.P. (1983) "Lan9uage maintenance, shift and deafh and the implicat1ons for bilingual education" Notes on Literacy, 39: 10-2 1. 146 McCONVELL , P. (1985 � "Domains and code-switching among bilingual Abor�gines". In Clyne (ed. ), 1985: 95-125. McCONVELL , P. (1986) "Aboriginal language programmes and l ngu ge maintenance" . Australian Review of Applied � � . L�ngu�st�cs, 3: 108-122 . McCONVELL , P. (1991a) "Understanding language shift : a step towards language maintenance". In Romaine (ed. ), 1991. McCONVELL, P. (1991b ) "Cultural domain separation: two-way street or blind alley? Stephen Harris and the neo-Whorfians on Aboriginal Education". Australian Aboriginal Studies, 1: 13-24 . McBLHANON , K.A. (ed. ) (1975) Tok Pisin I Go We? Ukarumpa: Linguistic Society of PNG. MILROY , L, (1982 ) "Language and group identity" . Journal of Multilingual and Multicultural Development, 3: 207-16. MILROY , L. (1987a) Language and Social Networks (2nd ed . ). Oxford: Basil Blackwell. MILROY , L. (1987b) Observing and Analysing Natural Language. Oxford: Basil Blackwell. MILROY , J. & MILROY , L. (1985 ) "Linguistic change, social network and speaker innovation". Journal of Linguistics, 21: 339-384 . MILROY , L. & MILROY , J. (1992) "Social network and social class: toward an integrated sociolinguistic model" . Language in Society, 21: 1-26. MUHLBAUSLER, P. (1979) Growth and Structure of the Lexicon of New Guinea Pidgin. Canberra: ANU Pacific Linguistics Series, C-52 . MUHLHAUSLER, P. (1992 ) "Preserving languages or language ecologies? A top-down approach to language survival" . Oceanic Linguistics, 31: 163-180. 147 MYERS-SC?TTON, C. (1993a) "Common and uncommon ground : soc1al and structural factors in code-switching". Language in Society, 22: 475-503. MYERS-SCOTTON, C. (1993b) Social Motivations for Codeswi tching: Evidence from Africa . Oxford: Clarendon Press. MYERS-SCOTTON , C. (1993c ) Duelling Languages: Gramma tical Structure in Codeswitching. Oxford : Clarendon Press. NATIONAL LITERACY & AWARENESS COUNCIL (1989) A Language and Literacy Policy for Papua New Guinea . Port Moresby : Department of Education . NATIONAL STATISTICS OFFICE (1985) 1980 National Population Census. Port Moresby : National Statistics Office . NATIONAL STATISTICS OFFICE (1990} 1990 National Population Census . Unoffic1al results obtained by personal communication. NEKITEL, 0. (1992 ) "Culture change, language change : the case of Abu ' Arapesh, Sandaun Province, Papua New Guinea". In Dutton et al (ed).), 1992 : 49-58 . OLSSON , M. (1984 ) "Merging formal with informal through vernacular education". Yagl-Ambu , 11: 39-51. PARASHER, S.N. (1980) "Mother-tongue-English diglossia: a case study of educated Indian bilinguals ' language use". Anthropological Linguistics, 22(4): 151-68. PAWLEY, A. (1992 ) "Kalam pandanus lan9uage � an old New Guinea experiment in language eng1neer1ng" In Dutton et al . (eds.), 1992 : 313-334 . PNG DEPARTMENT OF EDUCATION (1991) Educational Sector Review. Port Moresby. RENCK G. (1990 ) Contextualisation of Christianity. Erlangen: Evangelisch-Lutherische Mission. 148 RENCK, G. (1992 ) "Values and semantic changes in Tagaria, Eastern Highlands Province, Papua New Guinea" . In Dutton et al (ed) .), 1992: 59-66 . RICHARDS , J., PLATT, J. & WEBER, H. (1985) Longman Dict�o. nary of Applied Linguistics. Essex: Longman . ROGERS, E.M. & KINCAID, D.L. (1981) Communication Networks : Toward a New Paradigm for Research . New York: The Free Press. RYAN , E.B. (1979) "Why do low-prestige language varieties persist?". In Giles & St Clair ( eds. ) , 1979, 145-157. SANKOFF, G. (1976) "Political J?OWer and linguistic inequality in Papua New Gu1.nea" In Sankoff, 1980. SANKOFF, G. (1980) The Social Life of Language . Philadelphia : University of Pennsylvania Press. SAOVANA-SPRIGGS, R. (1992 ) "Maintaining the langua<1es of the North Solomons : the Viles Tokples Skul ProJect". In Dutton et al (ed) .), 1992: 645-653. SAVILLE-TROIKE, M. (1982 ) The Ethnography of Communication; An Introduction. Oxford: Basil Blackwell. SCHLIEBEN-LANGE , B. (1977 ) "The language situation in Southern France ". International Journal of the Sociology of Language, 12: 101-8. SCHMIDT, A. (1985) Young People's Dj irbal . Cambridge: Cambridge Un1.versity Press. SCHMIDT, A. (1990 ) The Loss of Australia 's Aboriginal Language Heri tage . Canberra: Aboriginal Studies Press. SCHOOLING, S. (1990) Language Maintenance in Melanesia . Dallas : Summer Institute of Linguistics. SCOTTON, C.M. (1982 ) ' The p ssibili�y of sw tc�ing: ,' . ? . � motivation for ma1.nta1.n1.ng mult1.l1.nguall.sm • Anthropological Linguistics, 24: 432-444. 149 SCOTTON , C.M. (1988 ) "Code-switching as indexical of social negotiations". In Heller (ed. ), 1988 . ( 1977 ) SCOTTON , C .M. & U�Y, W. . "Bi�ingual strategies: the soc1al. funct1on of code-sw1tch1ng". International Journal of the Sociology of Language, 13: 5-20. SEBEOK , T. (ed.) (1960) Style in Language. Wiley. SELIGMAN , C.G. (1904 ) The Melanesians of British New Guinea . Cambridge : Cambridge University Press. SENFT, G. (1992) "As time goes by ••. : changes observed in Trobriand Islanders' culture and language , Milne Bay Province, Papua New Guinea". In Dutton et al (ed).), 1992 : 67-89 . SIEGEL, J. (1992) "The future of Tok Pisin: another look" . In Dutton et al. (eds. ), 1992 : 405-408. SMITH, G. (1990) "Creolized Tok Pisin: uniformity and variation". In Baldauf & Luke (eds.), 1990: 197-209 . SMITH, G. (1992 ) "Language obsolescence in Morobe Province, Papua New Guinea : two contrasting case studies". In Dutton et al (ed).), 1992: 115-121. SMITH , G. (1995 ) Personal communication by e-mail, 6th February . STRINGER, M.D. & FARACLAS , N.G. (1987 ) Working Together for Literacy. Wewak: Christian Books Melanesia. SWATRIDGE, C. (1985) Delivering the Goods : Education as / Cargo in Papua New Guinea . Melbourne: Melbourne University Press. TABOURET-KELLER, A. (1968) "Sociolinguistic factors of language maintenance and language shift : a methodological a�proach based on European and African examples". In F1shman et al . (eds. ), 1968, 107-118. THURSTON, W.R. (1992 ) "Sociolinguistic typology and other factors effecting change in north-western New Britain, Papua New Guinea" . In Dutton et al (ed) .), 1992 : 123-139. 150 TRUDGILL, P. (1983) On Dialect : Social and Geographical Perspect�ves. Oxford: Basil Blackwell. UNESCO (1953) The Use of Vernacular Languages in Education. Paris : UNESCO. VAN BOUT, R. & MUNSTERMAMNN , H. (1988) "The multidimensionalty of domain configurations". International Journal of the Sociology of Language , 74: 107-124 . WARDBAUGB, R. (1986) An Introduction to Sociolinguistics . Oxford : Basil Blackwell. WIERSMA, W. (1986) Research Methods in Education : An Introduction . Massachusetts : Allyn & Bacon (4th edition ). WURM, S.A. (1964 ) "Motu and Police Motu : a study in typological contrasts". Papers in New Guinea Linguistics No 2. Canberra: ANU . WURM, S.A. (ed. ) (1976 ) New Guinea area lan�uages and language study, 2. Canberra: ANU Pacif�c Linguistics Series, C-39 . WURM, S.A. (ed. ) (1977) New Guinea area languages and lan�uage study, 3. Canberra: ANU Pacific Linguistics Ser1es, C- 140. WURM, S.A. (ed. ) (1977 ) New Guinea and Neighbouring Areas: A Sociolinguistic Laboratory. The Hague: Mouton . 151 Appendix 1 QUESTIONNAIRE OH LANGUAGE USB IH BANUABADA 1. Place of Origin ------ 2 Clan (iduhu ) name 3. Age: 13-20 yrs 21-35 yrs 36-50 yrs over 51 yrs 4. Sex : M F 5. Religion (church group ) name 6. School attended/attending ------Grade 7. Do you or any of your family members work for an income? YES NO If yes, give details (who and what they do ) 8. What language/a do you speak at home ie in your family? 1 ------Why? 2 ------Why? 9. What language/a do/did you speak at school? 1 ------Why? 2 ------Why? 152 10. At work, what language/a do/did you use? 1 W�? 2 ------Why? 11. When you are in the city, what language/a do you speak? 1 Why? 2 ------Why? 12. What language/a do you or can you write? Bow did you learn to write this/these? 13. What language/a do you or can you read? Bow did you learn to read this/these? 153 14. Is Motu Gado of importance to you? YES NO If YES, how important is it? VERY IMPORTANT FAIRLY IMPORTANT LITTLE IMPORTANT 15. Give reason/s for your answer to question 14 . 16. Are any of the other lan9uages you know of the same importance as Motu Gado 1s to you? YES NO If yes, which languages and why? 1 Why? 2 Why? 17. Are there any other languages which you think it is (or will be ) important for you to know? 1 ------Why? 2 Why? 18. What language/a should children learn at home? 1 ------Why? 2 Why? 154 19. What language/s should children learn at school? 1 ------Why? 2 ------Why? 20. In what language should children be taught at: Grade 1 Why? Grade 2 Why? Grade 3 Why? Grade 4 Why? Grade 5 Why? Grade 6 Why? High School Why? Tertiary Why? 21. Other comments: 155 Appendix 2 RESPONDENT CHARACTERISTICS Ho. Orig Clan Age Sex Church Sc. Worker Lang of Literacy Grade M B TP 0 01 B B c M u 6 F,B,S T 02 B A B T T T M u 8 s ST T 03 8 B B M c 8 8 F,M,S T 04 M D M u 10 OS B A c T T ST M u 6 ST T 06 B 8 E M u 3 B R w w 07 M D M u 10 R T 08 B E T G M u 3 R ST ,W W,T 09 M 0 c M u 12 B R T ST 10 G D M u 10 R ST 8 E T ST 11 G M u 6 Sn ,D T T 12 B B c M u 10 B w T T 13 G B M c 8 M ST T 14 B M D M u 10 R B ST T ST 15 B c M u 6 W,R T,ST T 16 B B B M c 6 A,U B ST T 17 T B M u 10 M,F T 18 B T c M u 10 M,R ST B T 19 M c M c 10 B,S ST T 20 B K D F u 10 H ST T B 0 ST 21 c F c 10 H,S ST T ST ST 22 B B c F c 10 ST T 23 B G E F u 4 Gs ,Gd T T 24 8 G D F u 8 R ST T 25 8 K E F u 5 S,D ST T 26 B w c F c 12 R ST T 27 B w E F u 5 Gs T T 28 K 0 D F u 10 B ST T - ST 29 B M c F u 10 H ST T 30 B B B F c 6 ST T 31 B 8 c F c 6 F,S T 32 B 8 c F u 9 H ST T 33 8 B c F c 10 R,H ST T 34 8 G D F u 9 H ST T 35 8 T c F c 8 M,H ST T 36 8 A D F u 9 H T T 37 8 T E F u 6 Sn(4) ST T 38 B T D M u 8 Bl ST T 39 B w D M u 10 Sn T T 40 B w D F u 6 Sn ST T 41 B w 0 M c Un w T 42 B A 0 F u 9 ST T 43 B A E M u 9 0 ST T w 44 B A 0 F u 9 0 ST T 45 B A c F 0 8 H ST T 46 B A 0 M u 10 R ST T 47 B B 0 F c 9 R ST T 48 B B E F u 6 Gs ,N(3) T w 49 B B 0 M u 8 ST T 50 B B 0 M u 6 c T T 51 B G B M u 10 M.F ST T 52 B G 0 M u 11 B,N ST T w 53 B G c F u 10 R,S,M,B ST T ST 54 B G E M u 5 T T 55 s 0 c M c 10 R T T 156 Itey to table in Appendix 2:- Origins: Clans: Age: B - Banuabada A = Apau B = 13-20 = :K Kerema B = Botai c = 21-35 M = Misima G = Gun ina D = 36-50 s = = Sepik B Bohodae E = 51+ K = Kahanamona Worker: M = Mavara 0 = Outsider A = aunt T = Tubumaga B = brother w = Kwaradubuna D = daughter F = father Religion: Gs = grandson Gel = granddaughter c = Catholic B = husband u = United Church M = mother R = respondent Sex: s = sister M = male F = female Bow literacy in a particular language was aquired: ST • self-taught (e.g., by transfer of literacy skills) T = taught at school W • learned at work Col1U111 beadings No . • respondent number Orig • place of birth Sc . Grade • school grade completed by respondent M • Motu B • English TP • Tok Pisin 0 • other language 157 Apoendix 3 LANGUAGES USED IN MAIN SITUATIONS B ome s c h 0 0 1 Wor k c i t y Ho. M E 0 M E TP M E TP 0 M E TP 0 01 MT A M 0 LF LF 02 MT CL M CL LF LF 03 MT MT M I LF 04 v I M LF A 05 MT M LF LF LF LF 06 I s R CL LF CL LF 07 MT M 0 CL LF 08 8 M M 0 A LF 09 v F M I I LF LF 10 MT I CL M LF CL A A LF LF 11 MT A M M LF CL LF 12 MT CL M CL 0 A CL LF 13 MT I u M MT LF 14 u I LF 15 8 M LF CL LF LF 16 v X M LF 17 u LF 18 8 I,S - 0 LF 19 MT M 0 LF - CL 20 MT CL M CL A A CL LF LF 21 X X M CL LF LF 22 u A CL M LF LF LF 23 MT R 24 MT A M M CL LF 25 MT X c LF A CL LF 26 MT X M CL A CL LF 27 MT X M M CL LF 28 MT M CL 0 LF - c 0 LF 29 MT CL M LF LF - CL LF LF 30 v M CL LF 31 MT,8 - M CL LF LF - CL LF 32 MT,8 - M CL 0 CL 33 B M LF CL LF 34 MT M CL LF A CL 35 MT I,S - 0 LF 36 MT CL M LF LF 37 B MT I MT CL A 38 B M M CL 0 LF - CL A LF 39 MT M A A CL A A 40 MT M 0 CL A 41 c s M CL 0 LF - CL A LF 42 V,B CL M 0 LF - CL A LF 43 B CL M CL A A CL A A 44 MT ,B - CL M CL 0 CL A 45 B M 0 A CL A A 46 MT M M CL A A CL A A 47 MT X M O,A - CL A 48 B MT nc CL A CL A A 49 8 A M CL 0 A CL A A 50 MT X CL ,M M CL A A CL A A 51 MT M u LF 52 B,V M M 0 0 0 CL A A 53 V.MT s IU IU 54 B,MT M M CL A CL A 55 X X M LF LF - LF LF 158 �ey to tables in Appendices 3 & 4: Reasons given by respondents: A = for communication when non-Motuans are present B = bicultural & bilingual understanding c = cultural & language maintenance CL = communication link with fellow Motuans E = easier learning of other languages F = first language B = spoken by everyone in the horne I = improvement in the use of the language IU = international/universal language LF = common lan9uage or communication link with other groups M = medium of 1nstruction MT = mother tongue or native language 0 = official language, used for business R = used for religious or church purposes s = to assist communication at school or work ST = self-taught (e.g., by transfer of literacy skills) u = helps understanding v = spoken by everyone in the village X = parent/spouse/self speaks a language other than Motu nc = not categorised (because of unique or vague answer ) Col111111 beadings No . = respondent number M = Motu B = English TP .. Tok Pis in BM = Biri Motu 0 = other language VI - very imJ;>ortant FI - fairly 1..mportant 159 4 Appendix IMPORTANCE OF AND WHERE TO LEARN LANGUAGES Importance of Learning Learning Learning Motu other at Home at School in Language No. VI PI E TP M E M E M E TP BM 01 C,MT - IU s I I E 02 MT IU MT s s 03 MT s LF MT s s M 04 MT U I s M s 05 MT MT M 06 MT CL s M E 07 MT LF MT A A C,A 08 MT MT I 09 F LF LF CL IU IU LF 10 MT LF LF MT M M I 11 u MT M c c 12 R M LF MT M M I 13 nc I LF MT LF IU LF 14 MT LF s IU LF 15 CL LF M M M 16 nc M M 17 MT 18 MT LF LF s s s 19 MT MT M I 20 MT IU MT M M M 21 X MT M M LF 22 MT LF CL IU M LF 23 MT s s 24 MT I MT s E 0 I 25 MT M,E - MT s M M 26 B LF LF s 0 s 27 MT E s M I 28 MT M,I - MT M M 29 MT M MT M M M 30 MT IU u s U, I u,s u 31 MT MT s s s 32 MT LF MT M M 33 CL LF M M M 34 MT MT s M M 35 MT I s s s 36 MT MT s s 37 C,MT IU I IU M M 38 B,V, v MT MT 0 39 c I 0 0 0 0 40 c v, 0 0 0 0 41 c 0 I,E IU I,E IU 42 c M,O I s nc 43 CL u I LF 0 0 0 44 B nc nc 0 0 0 45 MT MT s S,IU S,IU - 46 c u, U,B U,B U,B U,B U,B 47 CL nc MT IU I 0 I 0 48 c IU c s B IU B IU 49 c M,O c s M M 50 CL ,C - c s,u U,B U,B U,B U,B 51 CL ,C LF c IU nc nc 52 c u u M c M c M 53 C,MT - u,s MT s MT ,U IU MT ,U IU 54 nc IU c B U,B U,B c u 55 v LF LF M,S M M 160 Appendix 5 SUMMARY OF REASONS FOR LANGUAGES USBD Home Lang School Lang Work Lang City Lang Used & Used & Used & Used & Reason Reason Reason Reason M B 0 M BTP M B TP 0 M BTP 0 A 5 10 12 1 16 7 1 c 1 1 1 CL 11 18 35 F 1 B 16 I 1 3 4 1 1 1 IU 1 1 LF 1 1 - 13 9 29 17 M 8 45 1 MT 32 3 1 1 0 - 18 1 1 1 R 2 s 3 2 u 3 1 1 v 7 X 7 3 nc 1 ltey to table : Reasons given by respondents : A = for communication when non-Motuans are present c - cultural & language maintenance CL = communication link with fellow Motuans F = first language B .. spoken by everyone in the home I - improvement in the use of the language IU = international/universal language - LF common lan9uagejcommunication link with other groups M - medium of �nstruction MT - mother tongue or native language 0 - official language , used for business R - used for religious or church purposes s - to assist communication at school or work u - helps understanding v - spoken by everyone in the village X = parent/spouse/self speaks a language other than Motu nc - not categorised (because of unique or vague answer ) Colu.n headings (languages used by respondents) : M • Motu E • English TP • Tok Pisin 0 • other language 161 Appendix 6 REASONS FOR IMPORTANCE AND LEARNING Importance of Learning Learning Learning Motu other at Home at School in Language VI FI B TP M B M B M B TP A 1 1 1 B 1 2 4 3 3 2 c 13 5 2 3 1 CL 4 2 3 B 2 1 1 2 F 1 B 2 1 1 I 3 3 4 10 IU 9 5 7 4 LF 13 7 1 1 5 M 7 11 21 13 MT 31 23 2 2 0 3 5 6 4 R 1 s 1 22 8 10 u 1 2 4 2 4 3 4 4 1 v 1 1 2 X 1 nc 1 2 1 1 3 1 ltey to table: Reasons given by respondents: A - for communication when non-Motuans are present B = bicultural & bilingual understanding c = cultural & language maintenance CL = communication link with fellow Motuans B = easier learning of other languages F - first language B .. spoken by everyone in the home I = improvement in the use of the language IU = international/universal language LF - common language/communication link with other groups M - medium of instruction MT = mother tongue or native language 0 = official lan�uage, used for business R - used for rel�gious or church purposes s = to assist communication at school or work u - helps understanding v - spoken by everyone in the village X - parent/spouse/self speaks a language other than Motu nc - not categorised (because of unique or vague answer) Cola.n beadings M • Motu B • English TP • Tok Pisin VI • very im�ortant PI • fairly �portant 162 Appendix 7 QUESTIONNAIRE ON LANGUAGE USB AND NETWORKS ...... RESPONDENT Is NAME: ...... clan: ...... origin : ...... .•••..••.••.•.... ••..•••. residence: age: sex: ...... •••••••••••..• school grade : religion: ...... work: ...... location: bred in village? attends ceremonies? spends .are than 50% spare time in village? •.••••••••••• Views on Tok Ples education : Agree Uncertain Disagree There should be Tok Ples pre-schools in Banuabada ...... 3 2 1 There should be Tok Ples bridging classes in Grade 1 3 2 1 Tok Ples literacy should be aaiDtaiDed i.n Caaauni ty School 3 2 1 Tok Ples should be used in non core subjects in Community School 3 2 1 Th ere should be Tok Ples .aintenance programmes at the 3 2 1 secondary and tertiary levels There should be Tok Ples proga..es for adults and • • • • • • • • 3 2 1 out-of-school youth 163 Which language do you use with the following pe opl e? grandparents: ...... parents: brothers/sisters: spouse: .•..•••.••••.•.•... grandchildren : own children : ...... friends : ...... neighbours : ...... traditional doctor: ...... modern doctor : ...... school children : ...... teachers : ...... village seller: ...... trade store assist: ...... .a.rket seller: ...... town sbop assistant : •••••• bank assistant: ...... government official: ...... church �nister: ...... God: ...... fel low worshipper: ••••••••• ...... supervisor : ...... · subordinate: . co-vorlcer: ...... 164 RBSPOHDBN'I' I s HAM.B: ...... Network Details: Con tact 's Contact 's Contact 's activities in the village: name network bred resides spends spends attends role working . 5 free ceremony day time ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... ...... 165 Appendix 8 CHARACTERISTICS OF TBB RESPONDENTS FOR PART 2 Eorly Attends Free Mo. Clan Origin Hoe Age Sex Grode Church Work life cerea'y tie 1 Kwa radubuna HB Kone H 30 12 u POM 0 y 0 2 Kwa radubuna HB Wewok H 26 8 c Wewak HB y HB 3 Mevara 0 HB H 25 10 c HB y HB 4 Botat HB HB 28 H 12 u POM HB y HB 5 Kehane.one HB HB H 42 9 u HB y HB 6 Outsiders HB/0 Soroo F 17 10 u 0 y HB 7 Hohodee HB HB H 30 10 u Boroko HB y HB 8 Mevara HB Toks f 32 10 u POM HB y HB 9 Hohodee HB Kone f 44 12 c Kone HB y 0 10 Hohodee HB HB 26 H 10 c Boroko HB y HB 11 Apau HB P<»> 26 f 12 u POM HB y 0 12 Gun ina HB P<»1 40 f 10 u POM HB y 0 13 Hohodee HB Gordons 26 F 13 c POM 0 y 0 14 Tubu•g• HB HB 42 H 6 u HB y HB 15 Botat HB HB 23 f 10 u HB y HB 16 Botat HB HB 31 M 13 u HB y HB 17 Gunine HB HB 28 F 5 u HB y HB Guntne 18 HB HB 59 F 8 u HB HB y HB A�u 19 HB Borakou 28 f 13 u POM HB H 0 Gunlne 20 0 HB 40 F 10 u HB y HB 21 Mevare HB HB 53 H 9 u HB y HB 22 Guntna HB HB 47 H 13 u HB y HB 23 Tubu•g• HB HB 44 H 8 u HB y HB 24 Guntne HB A4baul 35 H 11 u HB H 0 25 Hohodae HB HB 52 F 6 u HB y HB 26 Hohodae HB HB 69 M 3 u HB y HB 2 7 Apeu HB HB 35 f 9 u HB y HB 28 Apeu HB HB 52 H 6 u HB y HB 29 Klferadubune HB HB 38 M 13 c HB y HB 30 ltlfaradubune HB HB 60 F 6 u HB y HB 31 K.lhena.one HB HB 54 M 5 u HB y HB 32 K.lhene.one HB HB 30 H 12 u 0 N 0 33 Mevera HB Boroko 55 H 13 u Boroko HB y 0 34 Mevera HB HB 23 F 10 u HB y HB 35 Mev ere HB Soroo 27 F 8 u Saroa HB y HB 36 Botet HB Hoholo 26 f 12 u Gordons HB H 0 37 Botat HB HB 73 F 6 u HB y HB l8 BoteI HB HB 29 F 10 c POM HB y HB 39 Botet HB HB 26 F 8 u HB y HB 40 Guntna HB HB 49 M 13 u POM HB y HB 41 Gunlna HB HB 42 M 10 u HB y HB 42 Guntna HB HB 36 H 6 u HB y HB 43 Gunlne HB HB 26 F 10 u POM HB y HB 44 Botat HB HB 28 H 6 u 9 Hi le HB y HB 45 Hohodae HB P lfJ!y: HI • #Mnud•d• U • Un fted Church L • Ldtter Ddy Sdints 0 • Outl #de C • C•tho lfc Church 166 Appendix 9 LANGUAGES USED WITH VARIOUS INTERLOCUTORS Resp Language Used With Each Interlocutor Type Mo. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 01 M M M ME - ME MEP MEP - E ME E M M PH PH MEP MEP ME ME ME ME ME ME 02 M M M - MEP M - MEP ME MEP M M P EP EP EP E E ME EP EP EP 03 MD MD M - MEP M - MEP ME MEP M H MP MEP HEP HEP E E HE 04 M M M M - M MEP M - E ME E M H HP HEP HEP HEP H MEP HE EP MEP MEP 05 M M M M ME M HEP H - ME ME ME M M MEP HEP MEP MEP H M H 06 M MD 0 - - - HEO MD MD ME 0 E MO MO MO E E E MO0 07 M H M M - ME MEP M M E HE E M H HE E E E H HE M E E E 08 M M M M - ME ME ME M E E E M HE H E E E E HE ME E E E 09 M M H M - ME E E H E E E M H H E E E E E ME E E E 10 H M H - - - MEP H - MEP M E M H M ME E E E E M E p E 11 M H M ME - ME ME E M ME M ME M H HE HE E E H M H E E E 12 M M H M - M MEP EP M HEP M E M H HEP MEP E E HE H ME E E E 13 M M M - MEP MEP M ME HE MEP H H H E E E E E HE EP E E 14 M M M M M H H M H HEP H HE H H HE HE HE ME H H H 15 M M M M - M MEP H M MEP H ME H H HEP MEP HEP HE H HE H 16 M M M M - M MEP M M MEP M MEP H H MEP HEP HEP HEP H H M 17 M M M M - M MEP M - HE H HE H H HP HEP HEP HEP H H H 18 M M M M M M MD M M - ME M H H H HE HE HE M H H E M E 19 M M M ME - ME ME M M E HE HE H H H HE HE ME M M H E E E 20 M M M M M M ME H M ME M ME M H HE HE HE HE M H H 21 M M M M M M MEP M MH MEP M MEP H M MEP MEP MEP HEP M M M 22 M M M ME - M MEP MEP H ME MEP E M M HP HE E EP M H ME - Z3 M M M M M M MEP M M ME M MEP M M MEP MEP HEP HEP H H H 24 M M M M M MEP HEP EP - EP EP EP P P P EP EP P EP HEP HEP - 25 M M M M M M M M - HE M HE H H HP HEP HEP - H M M 26 MD M M M M M M M - EP H E M M HP HEP HEP MEP H H M 27 M M M E - ME ME M H ME H HP H H HEP HEP HEP MEP H H H Z8 M M M M M M MEP M M HEP HE HEP H M HP EP HEP EP HP H H 29 M M ME ME ME ME ME M M HE E HE M M HP HEP E E ME E M 30 M H M M M M M H H H H H H M H H H H M M M 31 M M M EO MEP ME MEP H - E ME E H M MP HE HEP MEP M M M 32 M ME ME EO - E MEP MEP H E E E H M HP MP HP HP HEP E MEP - 33 M M M M M M MEP MEP H MEP ME E H M P P E E ME H H E E E 34 M M M M - M MEP H M ME M MEP M M MP HE HEP HEP H H H 35 M M ME M - MEP MEP MEP H HE HE ME M M MEP HEP EP EP HEP HEP HEP HEP MEP HEP 36 M M M E - E ME P - ME E E H H H P E HE H E H EP EP EP 37 M M M H M M ME H M ME M ME H H M HE HE HE H H H l8 M M M M M H ME M M E M E H M HP HP ME E H HE H E E E 39 M M M 0 - EO HEP M - E M ME H H P E HEP ME H M H 40 M M M M M M MEP MEP M MEP ME MEP M MP MP MEP MEP E MEP MEP ME M M E 41 M M M M - M HEP M MH ME M MEP M H HEP MEP HEP HEP H M M 42 M M M M ME M MH ME M MEP M M MEP HEP MEP MEP M H M 43 M M M - - M MEP M M E E E M H M E E E M H H E E E 44 M M M M - M MEP H M E M E H MEP M E E E H H M E E E 45 - - M ME ME ME ME E M E E E M HE ME ME E E E ME ME E E E 46 - M M - ME - MEP M M E HE E M H M HE E HEP H M M E E E 47 M M M M - M HE HE - E ME ME M HE HE HE E E E E ME E HE E 48 M M H ME - ME ME HE H HE ME E H HE HE ME E E ME ME ME E E ME 49 M M M M - H HEP M M E M HEP M M H HEP EP EP M M M E MEP MEP 50 M M M M - ME E M M E E E M H M E E E M H M E E EP 51 ME M M M • M ME M M E ME E M M MP E E E H HE H E HE ME 52 M M M E - ME MEP M M E E E M M M E E E H ME M E E MEP 167 Key to Appendix 9 Colu.n headings: Hw.ber Interlocutor 1 grandparents 2 parents 3 siblings 4 spouse 5 grandchildren 6 own children 7 friends 8 neighbours 9 traditional doctor 10 modern doctor 11 school children 12 teachers 13 village seller 14 trade store assistant 15 market seller 16 town sho� assistant 17 bank ass�stant 18 government official 19 church minister 20 God 21 fellow worshipper 22 supervisor 23 subordinate 24 co-worker LanCJUages : M = Motu B = English p = Tok Pis in B = Biri Motu 0 = other language 168 Appendix 10 VILLAGBNESS Villageness Villageness S Resp c o r e R a n k No. Resp Netwk Resp Netwk 1 20 41 3.0 12.5 2 50 41 11.0 12 .5 3 100 77 41.0 36 .5 4 80 53 23.0 20.0 5 100 60 41.0 28.0 6 60 27 14 .5 3.0 7 80 59 23.0 26.5 8 60 42 14 .5 15.5 9 40 35 7.5 8.0 10 80 69 23.0 32 .0 11 40 25 7.5 2.0 12 40 39 7.5 10.0 13 20 30 3.0 4.0 14 100 96 41.0 51.0 15 100 77 41.0 36 .5 16 100 83 41.0 41.5 17 100 80 41.0 38.0 18 100 82 41.0 40.0 19 20 31 3.0 5.5 20 100 93 41.0 45.0 21 100 64 41.0 29.5 22 100 41 41.0 12.5 23 100 95 41.0 48.5 24 20 58 3.0 25.0 25 100 95 41.0 48.5 26 100 95 41.0 48.5 27 100 65 41.0 31.0 28 100 97 41.0 52.0 29 100 83 41.0 41.5 30 100 95 41.0 48.5 31 100 50 41.0 19.0 32 40 41 7.5 12 .5 33 50 15 11.0 1.0 34 100 93 41.0 45.0 35 80 54 23.0 21.5 36 20 33 3.0 7.0 37 100 93 41.0 45 .0 38 80 64 23.0 29 .5 39 100 43 41.0 17 .0 40 80 38 23.0 9.0 41 100 88 41.0 43.0 42 100 81 41.0 39.0 43 80 54 23.0 21.5 44 80 76 23.0 35 .0 45 50 45 11.0 18.0 46 80 57 23. 0 24 .0 47 80 71 23.0 33.5 48 60 56 14.5 23.0 49 80 59 23.0 26.5 50 80 71 23.0 33.5 51 80 42 23.0 15.5 52 60 31 14.5 5.5