Working out languages: An interactionist analysis of vitality issues for Taiwanese Austronesian languages

TitreWorking out languages: An interactionist analysis of vitality issues for Taiwanese Austronesian languages
Publication TypeChapitre d'ouvrage
Année de publication2006
AuthorsSaillard, Claire
Book TitleStreams Converging into an Ocean: Festschrift in Honor of Professor Paul Jen-Kuei Li on His 70th Birthday
Series VolumeLanguage and Linguistics Monograph Series
PublisherAcademia Sinica
ISBN Number978-986-00-6784-2

Claire Saillard

UFR Linguistique

Université Paris 7, France


Working out languages:

An interactionist analysis of vitality issues for Taiwanese Austronesian languages



This article presents the results of a field research in interactionist sociolinguistics conducted in Hualien. Through the description of multilingual strategies in face-to-face interactions between Austronesian language speakers and their co-workers in two given professional settings, the research aimed to answer the question of vitality of Austronesian languages in a multiethnic society like Taiwan. It was found that, although the workplaces investigated were favorable to the use of Austronesian languages both for professional and extra-professional purposes, Austronesian languages were not the preferred choice of the Aborigine workers. The prevalence of standard Chinese was made quite clear whenever the speakers mastered this language sufficiently. Other non-official languages, such as Japanese and Minnan Chinese, were found to fulfill some unexpected social functions. An investigation of language attitudes as indicated by the names given to Austronesian languages corroborated the feeling of linguistic insecurity exhibited by the Aborigine speakers’ language choices. The article ends with some perspectives on the future of Taiwanese Austronesian languages in a changing society.


Keywords : Interactionist sociolinguistics, Taiwanese Austronesian languages, language choice, language vitality, functions of languages.


Full Text

Claire Saillard

UFR Linguistique

Université Paris 7, France


Working out languages:

An interactionist analysis of vitality issues for Taiwanese Austronesian languages


1. Foreword

The development of graduate studies in linguistics in Taiwan since the late 1980s has been a formidable stepping stone for the development of Taiwan’s Austronesian languages studies, as can be seen from the wealth of scientific literature on the subject nowadays, ranging from Master’s theses to full-bodied research project reports and academic publications. My choice of a Doctorate research subject in the early 1990s was directly linked to the stimulating novelty that was offered to us in Tsinghua University through Dr Paul J.-K. Li’s graduate course on “Field methods”. My first year in Tsinghua’s Graduate Institute of Linguistics had given me a strong taste for Taiwan’s autochthonous Chinese languages, but the introduction to Austronesian languages through fieldwork opened for me a realm of questions that extended geographically to the whole Asia-Pacific area, and conceptually to issues ranging from historical linguistics to sociolinguistics and language policies. The article presented here is based on fieldwork dating from 1996 through 1997, which is to say the facts described here may already belong to the past, and the analysis developed at that time may not be supported by the 2006 facts. But I feel compelled to offer this article in honour of Dr Paul J.-K. Li for at least two reasons. First, though I have published in French on several aspects of this research, no paper was published either in English or in Chinese after the completion of my dissertation (Saillard 1998a)[1]. Thus, the present paper is a great opportunity to inform the people whom I worked with “in the field” about the results of a two-year long research. This article may also enable my Taiwanese mentors, especially Dr Paul J.-K. Li, to see what has become of the questions their courses raised in their student’s mind. Second, although recent year have seen a wealth of research publications on the macro-sociolinguistics of the languages spoken in Taiwan, including Austronesian languages, there are to my knowledge no published studies on the latter based on an interactionist sociolinguistic analysis. In this article, it will be argued that an analysis of the mechanisms of multilingual interactions provides new insights in the question of the future of Taiwan’s minority languages.


2. Introduction: Why investigate multilingual interactions at work?

Interactionist analyses of multilingual language uses could be performed in any setting, and indeed could be performed most easily in informal settings such as the home and neighborhood. Why then choose to investigate language choice mechanisms in the workplace?


2.1. Language in the workplace

First of all, why is language at work significant in the case of Taiwanese Austronesian languages? In multilingual societies, Fishman (esp. 1964, 1967)[2] has shown that the functional distribution of language uses guarantees the stability of linguistic situations in terms of the economy of language resources. That is to say, any language that will not answer specific purposes is doomed to oblivion, because it will be replaced by another language with those functions. As regards the Taiwanese situation, standard Chinese (Guoyu) has acquired functions over the past decades that were heretofore the domain of local languages, be they Chinese or Austronesian. As a consequence, local languages have seen their domains of use shrink, some languages being ultimately restricted to domestic communication, and excluded form socially more rewarding domains. In the 1990s, modernization was often cited as a major cause of local languages decline in Taiwan, and particularly the decline of Austronesian languages. It was argued that the Aborigines had been forced into a modern, multiethnic, and moreover, work-centered reality. Workplaces in Taiwan are mostly multiethnic, but not necessarily multilingual. That is to say, language use in the workplace is not regulated by law[3], but rather subject to more or less covert negotiation. It has been shown that all negotiations aiming to define individual identity are ultimately meant to help one social group or another access to prosperity, or even social control. Thus, the issue of those negotiations depends heavily on each group’s vitality at the local or national level. That is why in most cases in Taiwan, standard Chinese or Minnan Chinese prevail in the workplace. Furthermore, language choice and vitality are linked together through a circle that can be either virtuous or vicious: the more a language is used in the workplace, the more it will add to the corresponding group’s social vitality. This being said, choosing the workplace in order to investigate linguistic uses was motivated by the belief that work-related interactions are observable mechanisms of interethnic confrontations linked to social control strategies. Moreover, the outcome of language choice negotiations in the workplace is significant as regards more general interethnic relationships. Let us say a word here about attitudes and motivations. Speakers’ attitudes regarding languages they speak also determine language vitality, be it on instrumental or emotional grounds. The tendency in the 1990s was for Taiwanese minority languages’ instrumental value to decline, as a logical consequence of the functional atrophy that characterized those languages. Then the question must be asked whether speakers’ emotional motivation regarding their own ethnic language has a real impact on their practice.


2.2. Language choices as interactive mechanisms

Second, why choose to investigate language choice mechanisms?

Researches on Taiwanese Austronesian languages vitality published in the 1990s were mostly based on statistical studies of reported language use[4], and aimed mainly at quantifying speakers of a given language in contrast to members of a given ethnic group. Since most individuals in Taiwan are in fact speakers of several languages, each of those languages performing distinct functions, my research on the contrary was based on a qualitative approach, aiming to unveil mechanisms that prevail in multilingual practices. Those mechanisms were found to be meaningful not only to analyze ongoing micro-sociolinguistic situations, but also in a long-term, macro-sociolinguistic perspective.

Whereas Fishman’s theory about domain of language use (as synthesized in Fishman 1972) was meant to analyze language choice at the social level, further research (pioneered by Blom and Gumperz 1972) focused on social meaning as expressed through linguistic structure, in a micro-sociolinguistic perspective. Blom and Gumperz insist on the necessity to devise analyses at the speech event level, in order to understand what motivates the alternate use of languages in multilingual situations. This field of research has proven very fertile, and the strategic dimension of language choice[5] in its many forms (plain language choice, code switching, code mixing and spontaneous borrowing) has been ascertained by such models as the Communication Accommodation Theory (Giles et al., 1987; Giles and Coupland, 1991) or the Markedness theory (Myers-Scotton, 1993).


3. Methodology

3.1. Choice of the research fields

Since my focus was on Austronesian languages at work, it was necessary to find workplaces where homogeneous groups of Austronesian languages speakers were employed together. I figured out that Hualien would provide many opportunities of finding such workplaces. The Hualien district is home to the ‘Amis, Taroko (a branch of Sediq) and Bunun ethnolinguistic groups; 23% of its total population is of Aboriginal descent, but 5% only of its urban population is Austronesian. As compared to the 1.7% of Austronesian population in the whole of Taiwan, the Aborigine population of Hualien is very dense. However, finding the right workplace to conduct fieldwork proved difficult, not because there were no such workplaces, but because most the nature of the work itself prevented observation by outsiders while seriously limiting language interaction between coworkers[6].

I finally settled for two distinct workplaces, the first being the Central Supplies Room (CSR) of a private hospital in the northern part of Hualien City. The second was a Public Care Unit (PCU) and its network of nine Public Care Stations (PCS) in Hsiulin County, north of Hualien City. Both were chosen because they had a high rate of local aborigine employees. Given their size, they range as two of the most prominent units in Hualien’s public health network.


3.2. Characteristics of the two workplaces

3.2.1. Linguistic repertoires of the workers

Although the proportion of Aborigine employees is very high in the whole hospital, compared to other Hualien hospitals, the CSR was chosen because of its even higher percentage of Aborigine employees. Among the 11 employees working there between February 1996 and April 1997, 5 belonged to the 'Amis group, 4 were Taroko, 1 Hakka and 1 Minnan. The five ‘Amis employees were over fifty years old, while the Taroko employees were in their twenties, except for the supervisor, a middle-aged woman.

The array of languages spoken by the employees is quite wide. All are able to speak standard Chinese and their mother tongue, and a varying amount  of Minnan Chinese (only the Minnan and the Hakka employees are fluent Minnan Chinese speakers though). The older employees can speak Japanese in addition to the former languages. All use a fair amount of English loan-words either for professional of non-professional purposes. What is striking is that people belonging to the 'Amis group cannot speak Taroko, and vice versa, apart from a very limited number of words used formulaicly. This is to say that contrary to standard Chinese, Japanese, and even Minnan Chinese to a certain extent, Austronesian languages are used exclusively as group-internal languages.

In the Hsiulin PCU, all employees are likewise multilingual speakers. At the time of observation (May-August 1997), there were 23 employees, among which 12 Taroko, 3 Atayal, 2 ‘Amis, 3 Minnan, 2 Mainlanders and 1 Hakka. The minimum language repertoire contains standard Chinese and the mother tongue (if different). Nevertheless, some PCU employees’ linguistic repertoire is much wider than that of the hospital CSR employees. For instance, the most extended repertoire (Taroko—mother tongue, standard Chinese, Minnan Chinese, Atayal, Japanese) belongs to the medical Doctor in charge of the PCU. Similarly, an Atayal nurse boasts a five-language repertoire (Atayal—mother tongue, standard Chinese, Minnan Chinese, Taroko, Japanese), but has mostly receptive skills in the last two languages. Contrary to what prevailed in the hospital CSR, a large part of Aborigine employees (6/17) in the Hsiulin PCU master Minnan Chinese, whereas fewer are able to use Japanese, and only one (the medical Doctor) is able to use it for professional purposes[7]. Among the six Aboriginal employees able to use Minnan Chinese at work, five were trained in medical or nursing schools in the South-West and South parts of Taiwan, where Minnan Chinese is a dominant language. Most nurses who do not use Minnan Chinese at work say they understand it a little, because they have been trained in Minnan-speaking settings; however, as they were meant to work in “the mountains”, they didn’t want to invest in learning Minnan Chinese. All the Atayal employees have learned the Taroko language (considered quite similar by them to their native tongue), but master it to varying degrees.

As can be seen, linguistic repertoires of the Hsiulin PCU employees range from two to four types of languages (native language, official language, vehicular language, minority language[8]), but a fair half (12/23, among which 9 Taroko) only know their mother tongue and the official language. Given the fact that Hsiulin is populated by an overwhelming majority of Taroko people (see Table 4 in appendix), these two languages enable them to cope with most of their professional encounters.

3.2.2. Linguistic repertoires of the Aborigine workers’ interlocutors

In the case of the hospital CSR, most of the professional interactions take place either  internally in the CSR or with other hospital employees (warehouse workers, cleaning and maintenance workers, administration workers, nurses, exceptionally patients and visitors). Whereas there are a number of Austronesian language speakers among the hospital employees, they tend to perform less qualified jobs than Han employees (they are less represented among the doctors and nurses than among the cleaning and maintenance workers for instance). As a consequence, interactions on professional topics outside of the CSR tend to take place with the Han speakers.

As for the Hsiulin PCU and its nine mountain stations, professional interactions take place both internally and with the patients. Most patients are local elderly people or children. The former speak only their native tongue and a little Japanese. If their native tongue is neither Taroko nor Atayal, the Japanese language is the only means of communication. As for children, they learn standard Chinese quite early, and are poor Austronesian language speakers as a whole. In the locality of Chongde, there is a high proportion of Minnan population. They elderly Minnan people speak Taroko as well as their native tongue (this is not the case with the younger Minnan inhabitants).

Whereas it is always possible in the hospital to speak standard Chinese for professional purposes, this is not the case in Hsiulin, given that the elderly patients tend not to master Chinese. As a consequence, the professional environment in Hsiulin PCU is more favorable to Austronesian language use than the environment in the hospital.


3.3. Observation protocol

I was admitted as a volunteer worker in the hospital to conduct my fieldwork in the CSR from February 1996 to February 1997. As such, I was able to do both unqualified work and linguistic fieldwork through participant observation. The work activities I was able to participate in while observing linguistic interactions range from work in common (packaging, sterilization, inventory), rounds through medical units, service at the counter, to meetings of various kinds. Part of these activities was audio-taped, but the noise associated with packaging and sterilizing processes often interfered with the taping. At the end of the fieldwork period, I conducted interviews in standard Chinese with each worker about their language habits and attitudes.

The fieldwork in Hsiulin was conducted from May to August, 1997. It was impossible for me to work in that setting, so that I was more of an observer. However, growing familiarity with the employees allowed me to participate in certain administrative tasks during consultations in some of  the mountain stations. The kind of tasks I was able to observe, both in the PCU and the PCS, ranged from consultation and medical care, vaccination in schools, visits to sick patients, administrative paperwork and office work, to various meetings. Some of these tasks were audio taped. As was the case in the hospital, interviews with each worker about their language habits and attitudes were conducted at the end of the study.

During fieldwork, for every interaction I witnessed, I wrote down such particulars as the identity of the speakers, the language(s) used, the topic and length of the interaction, who initiated the interaction, and who initiated language switch if any. Taped interactions which exhibited code switching or mixing were transcribed and double-checked with one of the speakers whenever possible.


4. Main results of the study

4.1. Language uses and functions

4.1.1. Overall tendencies

Observation of language choice (be it through choice of the interaction language as a whole, switching, mixing, or spontaneous borrowing) in both settings revealed general tendencies that obtained in both workplace in certain situations. It was found that interactions taking place at the hospital CSR can be compared internal communication in Hsiulin. Language choice in both cases shows standard Chinese to be almost compulsory in interactions between individuals belonging to different ethnolinguistic groups (counter-examples in the Hsiulin PCU are mentioned below). When speakers belong to the same group though, the use of standard Chinese is still frequent. In relation to interaction topic, standard Chinese comes as the most frequent choice for professional interactions, its frequency decreasing for non-professional topics (the decrease is sharper in the hospital CSR than in the Hsiulin PCU). The functions of standard Chinese can thus be described as “wider communication” and “professional communication” in those two settings. In both workplaces still, the choice of Austronesian languages is triggered by the speakers’ membership in the associated ethnolinguistic group, as well as by the non-professional character of the interaction. Thus, the functions associated to Austronesian languages are “identity/solidarity marker” while use of standard Chinese in similar situations would signify “neutrality” of interpersonal relationships.

Let us describe language use and choice in more detail for each workplace. In the hospital CSR, ‘Amis and Taroko workers have a significantly different behavior. ‘Amis workers use standard Chinese more frequently than the ‘Amis language as a whole. The ‘Amis language is used more frequently between employees aged over 50, and when the interaction bears on extra-professional topics. So it appears that the ‘Amis language, being dissociated from the youth and the profession, both “modern” values, crucially lacks vitality. As to the Taroko language, even though it is less used than standard Chinese as a whole, it is spoken both by young and middle-aged employees, for whatever topic. But it is mostly used together with standard Chinese, through code switching and mixing. This also points to the Chinese language’s growing use in Taroko speakers’ linguistic behavior. Language choice in the hospital CSR is summarized in table 1 below.


Table 1. Language choice in the hospital, when at least one CSR Aborigine employee is involved[9]

Interaction type ð



Speakers’ characteristics ò

Different ethnolinguistic group

Standard Chinese

(Minnan Chinese ; Japanese)

‘Amis group

Different age

Standard Chinese

No age difference (>50 years)

Standard Chinese



Taroko group

Taroko and Standard Chinese (code switching and mixing)


As stated in 3.2.2 above, interactions in the Hsiulin PCU differ greatly according to whether they involve only professionals or whether they involve patients. Between professionals (medical doctor, nurses, pharmacists, technical staff etc.), standard Chinese is used in most cases, not excluding code switching/mixing between standard Chinese and Taroko between Taroko and Atayal group members. With patients, the Taroko language is used dominantly only when an aged (monolingual) Taroko patient is involved. With middle-aged Taroko patients, code switching and mixing between Taroko and standard Chinese languages is the rule. With young Taroko patients or patients belonging to other ethnolinguistic groups, standard Chinese is the preferred choice. Japanese and/or Minnan Chinese can also be used with aged patients, when they understand neither standard Chinese nor Taroko. These tendencies are sketched in tables 2 and 3 below.


Table 2. Language choice in the Hsiulin PCU in interactions between professionals, when at least one Aborigine employee is involved

Interaction type ð



Speakers’ characteristics ò

All belong to the Taroko and/or Atayal groups

Standard Chinese (Taroko)

Standard Chinese and Taroko

One of the speakers is Han

Standard Chinese (Minnan Chinese)

Standard Chinese and Minnan Chinese

One of the speakers belongs to a different Austronesian ethnolinguistic group

Standard Chinese


It has to be stressed that when at least one Han employee is present, not only is standard Chinese used, but even Minnan Chinese can be chosen by the Taroko/Atayal group members. This peculiar language behavior will be commented below.

Language choice in interactions between professionals and patients is guided primarily by the patient’s characteristics (age, then ethnolinguistic group), then by the employee’s language repertoire, as shown in table 3. Local languages (be they Austronesian or Chinese) are considered by the PCU employees as professional tools, because it is necessary to accommodate to aged patients in order to communicate effectively. But using local languages is not compulsory; it rather is a voluntary choice on the part of the PCU employees, some of them even choosing to learn a local language different from their own.


Table 3. Language choice in the Hsiulin PCU in interactions between Aborigine employees and patients [10]

Interaction type ð



Patients’ age

Ethnolinguistic groups

Young patients

Any group

Standard Chinese

Standard Chinese

(local languages)

Aged patients

Taroko and/or Atayal groups

Taroko ;

(Standard Chinese)




Standard Chinese;

Minnan Chinese

Minnan Chinese;

Standard Chinese


Standard Chinese

Other group

Standard Chinese;

Japanese; ‘Amis.

Japanese; ‘Amis;

Standard Chinese.


4.1.2. Peculiar functions associated with languages Japanese in the CSR

Part of the interactions using Japanese in the hospital CSR were uttered entirely in Japanese, some exhibited code switching/mixing with standard Chinese or ‘Amis, and the most part were only instances of spontaneous or formulaic borrowing. Remarkably, most took place exclusively between workers belonging to the ‘Amis group. This group-internal use is quite surprising, given the traditional view that Japanese is an exogenous vehicular language for old Taiwanese who do not share any endogenous language[11]. Likely enough given their education background, only older ‘Amis workers were able to use Japanese throughout a sentence, whereas younger ones only used formulae, like greetings and thanks. The recurrent use of Japanese between older ‘Amis employees in the hospital clearly fulfills a different function from that of the ‘Amis language, which is the less marked choice in that context. Whereas the ‘Amis language is the most natural identification marker for ‘Amis people, Japanese is special in that it belongs to the older generation, so that it defines a special subgroup inside the already existing ‘Amis group. Moreover, Japanese, contrary to the ‘Amis language, is not used for efficiency reasons, since the speakers’ Japanese language skills are in fact limited. If we follow Calvet’s definition of a “langue grégaire” (as opposed to a vehicular language, Calvet 1987: 79-80, 82), Japanese, even in its reduced linguistic form, is used in this context in order to draw a line between those who can speak it and those who can’t, and thus give the speakers a feeling of belonging to a peculiar group.

Two complementary explanations could be given for the choice of Japanese as a group-identification language in this context.

First of all, the Japanese language still retains prestige dating back from the time it was the official language in Taiwan, and the language through which access to education was granted. Even though Japanese was later replaced by standard Chinese for its official functions, it still is valued as a language spoken by a well-educated minority, and linked to a culture with great influence in Asia. Elderly employees in the hospital CSR, though some are not old enough to have attended school in Japanese for many years, are not as fluent in standard Chinese and well-educated as their younger colleagues. Thus, they can use the Japanese language to claim that, despite their apparent flaws, they belong to a prestigious minority.

The second possible explanation lies in the age-class system of the ‘Amis culture. Age-classes are designated by special names, and used to be marked by easily recognizable signs, such as different color of dress and distinct meeting houses. Thus, it is not surprising that elderly ‘Amis people taken out of their traditional surroundings would adopt class-identifying behaviors. It may be that the use of the Japanese language among elderly ‘Amis workers in the hospital aims at distinguishing this particular age-class from the younger employees. Minnan Chinese in the PCU

In the Hsiulin PCU, all Han Chinese employee, be they Minnan, Hakka or Mailanders, have a good command of Minnan Chinese. The language is routinely used for internal communication between Han employees, thus replacing standard Chinese as the language of the Han group in the PCU.

Moreover, it was found that almost half the interactions between Han and Austronesian workers used the Minnan language through code switching, mixing or spontaneous borrowing. Thus, the use of Minnan Chinese was clearly triggered by the presence of at least one Han participant. Extra-professional interaction topics were also found to trigger the use of Minnan Chinese more effectively than professional topics, as was expected for a non-official language. What is quite surprising is that more than half the utterances or switches in Minnan Chinese were initiated by Taroko or Atayal employees, and that Aborigine employees use spontaneous borrowings to Minnan Chinese even between themselves. Since Han employees in the PCU are a minority (only 6 Han, among which 3 Minnan, on a total of 23 employees), and since Hsiulin is a Taroko-dominated County, no situation could possibly be more favorable to the use of Austronesian languages, especially Taroko, in the professional domain. Nevertheless, Taroko and Atayal employees use the Taroko language only with those individuals who do no master standard Chinese. Moreover, whereas they could easily communicate with their Han coworkers by using only standard Chinese (an unmarked choice in this context), they accommodate to their Han colleagues’ linguistic conventions by using a marked code, namely Minnan Chinese, despite its usual group-internal values. That is to say the Taroko/Atayal employees consider Minnan Chinese to be more prestigious than standard Chinese in this given situation, and they accommodate “upward” (Giles et al. 1987) to their Han co-workers. This is an obvious sign of linguistic insecurity. It seems that because the Han workers use Minnan Chinese as a group-identifying language, the Aborigine workers use it in turn to show that they too identify to the same group. Since Minnan Chinese is growingly used in Taiwan to denote identification with “native” culture and values, it can be guessed that Austronesian employees in the PCU use it to stress their Taiwanese identity, and refuse exclusion from the Han group.


4.2. Stability of multilingual situations

The results summarized above show that different languages have distinct social functions in the workplace. Other things being equal, this should ensure the stability of multilingualism in the workplace, and the maintenance of minority languages in that setting. But several factors could endanger such stability.

As regards the inter-generation transmission of languages, it was mentioned earlier that use of the ‘Amis language in the hospital decreased among younger employees. As to minority language skills, all workers in both settings agree on the fact that younger employees are poorer speakers than their elder colleagues. Younger speakers know fewer words, they tend to use simplified morphology (especially morphology linked to predicates), they even tend to simplify the phonemic structure of the Austronesian languages.

Both findings corroborate Tsao’s study (Tsao 1997: 11-12) according to which there is a strong correlation between age and education level of Austronesian respondents on the one hand, and standard Chinese language skills vs. lack of skills in the mother tongue on the other hand. This can be explained by the gradual functional spread of standard Chinese over Taiwanese Austronesian languages, the latter getting to be used less frequently and for scarcer functions. That is to say standard Chinese has not only been added to the Austronesian speakers’ linguistic repertoire when it was needed to fulfill exogenous functions, such as those linked to education, extra-community work, media and exogenous cultural activities etc. Rather, standard Chinese is gradually being substituted to Austronesian languages, reaching to domains traditionally associated to community and family.

Finally, whereas it was found that local languages crucially answered professional functions in the Hsiulin PCU, all workers think that this will not be the case for many more years, since monolingual Austronesian language-speaking elderly people will eventually die, to be replaced first by their bilingual children, then by their monolingual standard Chinese-speaking grandchildren. Moreover, the Hsiulin PCU is one of the very few workplaces in Hualien where the use of Aboriginal languages still has an instrumental motivation[12].

Interviews with employees of both workplaces show that the decline in language use and the attrition in language forms is linked to a more profound insecurity, linked to a crisis of the Aboriginal identity. This topic will be addressed in part 5.

This does not amount to saying Austronesian language speakers have lost all motivation to use and to transmit their languages. Rather, their motivations are mainly emotional. But in the past few decades, instrumental motivations for language use and transmission have proved to overwhelm emotional motivations. Aboriginal parents favored standard Chinese language learning over Austronesian languages learning since early childhood. But recent awareness of the value of the Aborigine cultural heritage is turning the language transmission issue into a more political one, thus offering new emotional motivation to Austronesian language speakers.


5. More on linguistic insecurity

Language uses as observed in both workplaces indicate that Austronesian languages, though they could be used more freely in a quite favorable work environment, are de facto restricted to a few inconspicuous domains, as if overt use of minority languages when not necessary had a social cost for their speakers[13]. During the interviews, linguistic insecurity was made apparent from different angles, one of them being the avowedly poor transmission of languages to the younger generations, another more subtle angle being the names attributed by the interviewees to languages in general, and Taiwanese Austronesian languages in particular. I will focus on this last point.


5.1. Naming languages : a hierarchization process

In Chinese, naming a language entails classifying it into a hierarchical system, through a choice between at least three morphemes : wen , yu and hua . The first and third morphemes refer to the written vs. oral status of the language (wen means “writing” while hua designates “speech”, thus an unwritten language). As a rule, wen is used to designate languages with an official status and a written tradition, that is to say political and social hierarchy builds on the prestige of writing. The morpheme yu designates a (spoken) language, no mention being made as to the existence of a writing system associated to the language. Last, hua is often associated with names of places or peoples, and designates a language that has neither writing system nor social prestige, with a geographically or socially restricted realm.

As expected for non-written languages, Taiwanese Austronesion languages are most frequently designated by their speakers as “hua ”, as in the following example:

 (1) [ME-2 : 6] ‘Amis speaker, aged 64                                                                  


wo jiu  yong    ziji       de        hua

I then   use      self      Part.    speech

Then I use my own speech

In this example as in many others, the language itself is not named, only its status is given through the chosen morpheme, and the person (here ziji 自己) or the group (as in “我們原住民的話women yuanzhumin de hua”) to whom it is related. The same applies when naming languages of other Aboriginal groups :

(2) [ME-1 : 64] Taroko speaker, aged 43, about the ‘Amis language


tamen  xuyao  jiang    tamen  de        hua

they     need    speak   they     Part.    speech

‘They need to speak their speech’

In some cases, the interviewees do not even mention the term  hua , as if they hesitated on the status applicable to the language referred to:

 (3) [ME-6 : 9] Taroko speaker, aged 21                                                                 


xiang   amei    zu        de        women ye       hui       jiang    yixie    zanghua

like      ‘Amis group   Part.    we       also     can      speak   some   obscenity

 ‘As for the ‘Amis group’s, we also know some obscene words.’

In rare cases, Austronesian languages are termed as “yu ”, or even “yanyu 言語”:

(4) [ME-1 : 9-10] Taroko speaker, aged 43


haiyou zhendui           ta         jiang    de        shihou wo       jiu        hui      

then     toward she       speak   Part.    time     I          then     Aux.   

yong    women            de        yanyu

use       we                   Part.    language

‘Or else when I talk to her, I use our language’

As for the wen morpheme, it was only used in the interviews to designate a very prestigious exogenous language:

 (5) [ME-4 : 6] ‘Amis speaker, aged 52


haiyou shi       jiang    yixie    yingwen                      jiushi   jiandan                        de

then     be        speak   some   English-language       that-is  simple             Part.

‘Then (I) speak some English, that is simple (phrases)’


5.2. Naming languages: a categorization process

In many cases, languages are referred to via a category name, such as muyu 母語 (mother tongue), guoyu 國語 (national language), fangyan 方言 (dialect or local language). Nevertheless, what the speaker has in mind is not the category as such, but a given language, and his meaning is totally clear to the interlocutors. Let us investigate a very meaningful category for minority language speakers: the mother tongue category.

The use of the term muyu 母語 (mother tongue) can be exemplified as below:

(6) [XE-18 : 32] Taroko speaker, aged 45


wo       xiaohaizi         dou      bu        hui       jiang    muyu              a

I           child                all        Neg.    can      speak   mother-tongue FPart.

‘None of my children can speak (the/their) mother tongue’

This sentence seems  paradoxical, since every individual, by definition, speaks his/her mother tongue. Obviously, what this interviewee means is that her children can’t speak Taroko, which is their mother’s first language.  These children’s first language is standard Chinese, but it will never be referred to as their « mother tongue » in this context (especially when the term muyu 母語 is not qualified). The same kind of use of the term muyu 母語 is found in ready-made expressions such as muyu jiaoxue 母語教學, where muyu 母語 refers to a local language, which has to be taught to children through education, since it was not in most cases transmitted through their parents.

This generalized use of the term muyu 母語to mean “local language” is even more obvious when used by speakers of different languages, as below:

 (7) [XE-7 : 16] Atayal speaker, aged 35


na        zhebian           de        muyu                          shi       zai       xue

then     here                 Part.    mother-tongue            be        Aux.    learn

‘As for this place’s mother tongue, (I) am learning (it)’

In this sentence, the language referred to is obviously not the speaker’s mother tongue, but rather the local language, Taroko.


5.3. Naming languages : an identification process

The interviews led in both workplaces show clearly how difficult it is for the Austronesian language speakers to name not only languages, but more fundamentally, to name the people according to their ethnolinguistic group when they speak Chinese. There are indeed official terms in standard Chinese to designate Aboriginal groups and related languages, but the list of official terms does not really match the needs of Taiwan Aborigines, since many groups or languages have no name in Chinese[14]. There are a number of different ways to name those groups or languages, either through names of categories as seen earlier, or by using names of related groups or languages. Let us take the Taroko group/language as an example. The Taroko group (truku in their own language) is considered a branch of the Sediq group, itself related to the Atayal group. At the time of the interviews, only the Atayal group was officially acknowledged as one of the nine “Mountain Aborigines” groups, despite linguistic evidence of the existence of a distinct Sediq branch. As a consequence, when they have to use the Chinese language, Taroko group member often refer to themselves as “Atayal”, rather than using the Chinese name tailuge 太魯閣, referring primarily to the Taroko Gorges and National Park, and sometimes used in the expression tailuge zu太魯閣族 to designate the group. During interviews, Taroko workers generally avoid the terms tailuge 太魯閣 and tailuge zu太魯閣族, but they visibly hesitate each time they have to name their group or their language. Their hesitation is marked in various ways. In the following example, the speaker first refers to the Atayal group through its official Chinese name, then switches to Taroko, using the real name of the language/group.

(8) [XE-18]  Taroko speaker, aged 45 (A), answering interviewer (Q).

Languages used : standard Chinese, TAROKO

1          Q         妳的母語是哪一個語言?

2          A         泰雅族的

3                      TRUKU


1          Q         ni de muyu shi na yige yuyan

2          A         taiyazu de

3                      TRUKU


1          Q         What is your mother tongue ?

2          A         The Atayal group’s

3                      TRUKU

The next example is a case of a phrase-internal code switching (or code mixing):

 (9) [ME-6 : 2] Taroko speaker, aged 21; languages used : TAROKO, standard Chinese

TRUKU -hua


‘The TRUKU-speech

The following example is more of a compromise, the name of the Taroko language being said neither in Chinese not in Taroko. It rather is an approximation of the official English name “Taroko”:

 (10) [ME-1 : 19-20] Taroko speaker, aged 43                                                                   


na.       women            dou      jia:ng   ei:        thaloko-hua

then     we                   all        spea:k er:        thaloko-speech

‘Well we all spo:ke er:  thaloko-speech’

In other cases still, all reference to the Taroko group is avoided, and replaced by a chimerical « Aborinal language » entity.

 (11) [ME-6] Taroko speaker, aged 21 (A), answering interviewer (Q)

3          Q         妳會說哪幾種語言?

4          A         好像只有國語跟那個:原住民

5                      這兩種

6          Q         原住民的哪一種?

7          A         就是:

8                      就是跟我們自己同族的


3          Q         ni hui shuo na ji zhong yuyan

4          A         haoxiang zhi you guoyu gen  nage: yuanzhumin

5                      zhe liang zhong

6          Q         yuanzhumin de na yi zhong

7          A         jiu shi:

8                      jiu shi gen women ziji tong zu de


3          Q         Which languages can you speak?

4          A         Seems like there’s only Mandarin Chinese and er: Aborigine

5                      These two

6          Q         Which sort of Aborigine ?

7          A         er we:ll

8                      Er well from the same group as us

In some cases, interviewees avoid referring to the language at all, naming only the associated group, with all the customary hesitation:

 (12) [XE-11] Taroko speaker, aged 40 (A), answering interviewer (Q)                           

1          Q         妳的母語是:哪一個語言?

2          A         泰雅

3                      就是太魯閣族


1          Q         ni de muyu shi: na yige yuyan

2          A         taiya

3                      jiu shi tailugezu


1          Q         What is your mother tongue ?

2          A         Atayal

3                      I mean the Taroko group

In all preceding examples, the ethnolinguistic group, the associated language and the corresponding individuals all seem to be interchangeable, showing how intricately group and language are linked for the definition of individual identity.


6. Conclusion

Since the 1990s, growing political awareness of the value of local languages for Taiwan’s cultural heritage and identity have led the government to encourage scholarly research and introduce local languages into school curricula, whereas most of the minority language conservation and teaching job had theretofore rested on local churches and communities, and the work of a few linguists. According to Fishman’s work on reversing language shift though, the school’s efficiency in maintaining minority languages is very limited in modern societies (Fishman 1991: 368-380), simply because children are socialized even before going to school. This is certainly the case for Taiwanese children, who learn standard Chinese through the media (if not through family members) before entering kindergarten.

However, Fishman pleads for measures focused not only on languages per se, but rather on restoring a certain kind of social fabric centered on family and community. Indeed, decline of community languages is but one of the visible consequences of a more profound social transformation (op. cit.: 4). This kind of social transformation that bears on family and community integrity is well documented for Taiwan Aborigines, who leave their villages in large numbers in quest for work in urban settings.

Paragraph 4.2 above stressed the fact that instrumental motivations for language use are currently stronger than emotional motivations, thus favoring official and majority languages over minority languages. Moreover, where instrumental motivations for the use of Austronesian languages still remain, like in Hsiulin, they seemed to be doomed to medium term extinction.

In parts 4 and 5, it has been shown through language uses and talk about languages that speakers of Austronesian languages such as Taroko face linguistic insecurity, if not a profound identity crisis. This certainly does not add to emotional motivations for language use and transmission.

This certainly is a negative conclusion, but not a desperate one. According to Fishman (op. cit.: 4-6), it is possible to revitalize any language provided that one defines reasonable and suitable objectives in the first place, and shifts to more ambitious goals gradually. Language revitalization policies could start by aiming at reinforcing emotional motivations for language use and learning, provided that the national context does not enforce competition between languages for economic, technical or political resources. At that point, the objective is to attain (or maintain) a diglossic functional repartition of languages, with local languages playing a valued role in local community and family-related domains.





Blom, Jan-Peter and John J. Gumperz. 1972. Social meaning in linguistic structure: code-switching in Norway. Directions in sociolinguistics. The ethnography of communication, ed. by John J. Gumperz and Dell Hymes, 407-433. Oxford: Basil Blackwell.


Calvet, Louis-Jean. 1987. La guerre des langues et les politiques linguistiques. Paris: Payot.


Caron, Jean. 1977. Stratégies discursives. Centre de Recherches Linguistiques et Sémiologiques de Lyon: Presses Universitaires de Lyon.


Ferguson, Charles A. 1959. Diglossia. Word 15(2): 325-340.


Fishman, Joshua A. 1964. Language maintenance and language shift as a field of inquiry. Linguistics 9: 32-70.


Fishman, Joshua A. 1967. Bilingualism with and without diglossia; diglossia with and without bilingualism. Journal of Social Issues 32: 29-38.


Fishman, Joshua A. 1968. Sociolinguistic perspective on the study of bilingualism. Linguistics 39: 21-49.


Fishman, Joshua A. 1972. Domains and the relationship between micro- and macrosociolinguistics. Directions in sociolinguistics. The ethnography of communication, ed. by John J. Gumperz and Dell Hymes, 435-453. London: Basil Blackwell.


Fishman, Joshua A. 1991. Reversing language shift. The theoretical and empirical foundations of assistance to threatened languages. Multilingual Matters.


Giles, Howard et al. 1987. Speech accomodation theory: the first decade and beyond, Communication yearbook, ed. by Margaret L. McLaughin., 13-48. Beverly Hills: Sage.


Giles, Howard and Nicolas Coupland. 1991. Language: Contexts and consequences. Mapping social psychology. Open University Press.


Han, Shih-fen韓世芬. 1996. 臺灣原住民語言能力及語言使用之調查研 A Survey of Language Ability and Language Use of the Aborigines on Taiwan. Hsinchu:  National Tsing Hua University M.A. thesis.


花蓮縣人口部. 1994a. 花蓮縣()平地山地原住民種族別人口數Hualian xian (shi) pingdi shandi Yuanzhumin zhongzubie renkoushu. 1222-91-01-2. Hualien District: Population Office.


花蓮縣人口部. 1994b. 花蓮縣鄉鎮市區十歲以上顯著就業人口數按性別及行業分Hualian xian xiangzhen shiqu shiwu sui yishang xianzhu jiuye renkoushu an xingbie ji hangye fen. 1222-03-01. Hualien District: Population Office.


Huang, Shuan-fan. 1993. Yuyan, shehui yu zuqunyishi -- Taiwan yuyanshehuixue de yanjiu. Taibei: The Crane.


Myers-Scotton, Carol. 1993. Social motivations for codeswitching, Evidence from Africa. Clarendon : Oxford Studies in Language Contact.


Robillard, Didier de. 1989. Vers une approche globale des rapports entre langue et économie.  Langues, Economie et développement, ed. by Robert Chaudenson et Didier de Robillard, 39-65. Paris: Didier Erudition.


Saillard, Claire. 1998a. Contact des langues à Taiwan : interactions et choix de langue en situations de travai. Paris : Université Paris 7 doctoral dissertation.


Saillard, Claire. 1998b. Quand variété veut rimer avec officialité, ou la situation sociolinguistique de Taiwan. Revue Française de Linguistique Appliquée III-1: 87-99.


Saillard, Claire. 2000a. Pratique et perception des langues minoritaires en situations de travail à Taiwan. Les langues en danger. Mémoires de la Société de Linguistique de Paris Nouvelle série VIII, 107-120. Peeters.


Saillard, Claire. 2000b, Nommer les langues en situation de plurilinguisme ou la revendication d'un statut (Le cas de Taiwan). Langage et Société 91: 35-57.


Saillard, Claire克來爾.  2002a. 言能力、使用與認: 蓮縣市田野調心得 Yuyan nengli, shiyong yu rentong: Hualian xianshi tianye diaocha xinde. 究通 10: 140-151.


Saillard, Claire. 2002b. Redéfinition des fonctions et valeurs symboliques des langues en milieu plurilingue, Paper presented at the Journées d'études internationales du Groupe d'étude sur le plurilinguisme européen. Langue et travail: champs et objets de recherche. Strasbourg: Université Marc Bloch.


Saillard, Claire. 2004. Les langues austronésiennes de Taiwan : peuvent-elles échapper à la minoration? Faits de langues. Les langues austronésiennes, 23-24: 361-378.


Tsao, Feng.-fu曹逢甫. 1997. 族群言政策海峡两岸的比 Zuqun yuyan zhengce haixialiang’an de bijiao. 台北 : 出版有限公司 Taipei: The Crane.


Zhang, Zhengliang 張政良. 1997. 社區健康評估實行成果Shequ jiankang pinggu shixing chengguo. 花蓮: 秀林衛生Hualien: Hsiulin Primary Care Unit.





Table 4. Aborigine population in Hualien District

(adapted from花蓮縣人口部 1994a)


Taroko (Atayal)



Total Aborigines

Hualien City


















































Xiulin (Hsiulin)















Hualien District (Total)








Table 5. Employment of adults over 15 years old

(adapted from Zhang 1997 : 7 花蓮縣人口部 1994b)

Job category

Hsiulin County (1996)

Hualien (1994)

Taiwan (1995)





Mines and quarries




Heavy industry




Water, energy




Light industry












Bank and services




Public services










[1] Saillard (2002a), the only article on this subject published in Taiwan and written in Chinese, was submitted in 1997, before completion of my research.

[2] Fishman’s findings were based on the pioneering work of Ferguson (1959) on the concept of « diglossia ».

[3] Contrary to language in the media, or in education.

[4] See the representative work of Huang (1993). Han (1996) and Tsao (1997), though based on a similar approach, include a focus on domains of language use.

[5] According to Caron (1977 : 181-185), for a phenomenon to be described as a strategy, it has to answer several criteria : (a) be part of an « uncertain situation », (b) have a goal, (c) obey rules, and (d) form a succession of choices. This can be shown to apply to language choice in interaction.

[6] Most workplaces in Hualien where Aborignal workers are more numerous than Han workers are construction sites, cement mines/factories and garbage-collecting brigades. See table 5 in appendix for employment trends in Hualien.

[7] In the hospital CSR, the older employees had elder siblings that went to school during the Japanese occupation. Similarly, some of the Hsiulin PCU employees have been raised under Japanese rule (not including the Doctor, who picked up his Japanese through practice with elderly Aborigine patients).

[8] Speakers who know five languages speak  two different minority languages, apart from their own.

[9] For  Tables 1, 2 and 3, languages are mentioned in frequency order ; languages mentioned between parentheses are barely used.

[10] The Taroko, Minnan and Japanese languages are used whenever the PCU employee is able to speak them ; otherwise, only Standard Chinese is spoken, no matter which group the patient belongs to.

[11] The Japanese language was taught and used for teaching other subjects throughout the Japanese occupation of Taiwan, which ended in 1945.

[12] Although there are PCUs in every county in the Hualien district, Hsiulin county is remarkably homogeneous compared to other counties. See Table 4 in the appendix section.

[13] In the interviews, some Taroko speakers report that the ‘Amis people are ashamed of their language and do not speak it in public, contrary to the Taroko people. Superficial observation of public places in Hualien seems to corroborate this opinion.

[14] The list of officially acknowledged Aborigine groups has been stable from 1954 to 2000. Since then, various groups have been claiming access to officialdom, on cultural and/or linguistic grounds.