Language Shift in Australia and Canada

J. K. Buda


Immigrants bring with them the language, culture, and religion of their country of origin. If the immigrants are either too few or too widely scattered to allow the formation of a viable minority community, maintenance of language, customs, traditions, and beliefs may prove difficult or impossible.

In the case of language, the social and economic necessity of using the official or majority language of the host country, and the lack of opportunities for using the mother tongue, may lead to a loss of ability in the latter. This loss of language ability, extended over several generations, will result in the phenomenon of language shift (or transfer), in which the habitual use of one language by a minority group is replaced by the habitual use of another. This shift to the second language usually, but not always, involves the gradual disappearance of the first.

For many years, the governments of multicultural English-speaking nations limited their interest in the phenomenon of language shift to an identification of the rate at which ethnic minorities were becoming fluent in English, and therefore of how much English language education might be necessary for these communities. The loss of the mother, or heritage, tongue was seen as a regrettable but unavoidable fact of life.

In recent years, however, the heritage tongues of ethnic communities have come to be viewed as a valuable and irreplaceable national resource, and official awareness of, and interest in, the phenomenon of language shift has grown.

This present study is an attempt to identify some of the major patterns of language shift in two multicultural nations: Australia and Canada.

It is based primarily upon the Australian Census of 1976 and the Canadian Census of 1986, with reference also to the results of earlier censuses, and to independent language surveys such as that conducted by the Australian Bureau of Statistics in 1983.

Most of the material for this study was collected during the course of a twelve-month study leave made possible by the generous support of both Otsuma Women's University and the Japan Association of Private Colleges and Universities.

Language Shift

The phenomenon of language shift takes place out of sight and out of mind. Individuals are rarely conscious of a subtle and gradual loss of language ability; still less are ethnic communities aware of the shift away from their heritage language and towards that of their host country.

From time to time the reality of this phenomenon is brought home in graphic form.

For example, an elderly matron approaches the counter at a police station in Warsaw and inquires about an amendment to her registration. She is dressed in typical rustic style—dark, shapeless, loose-fitting clothes and a colourful scarf wrapped tightly around her head. Her Polish is, however, halting and heavily accented. As personal identification she produces a United States passport.

An aged Indian gentleman enters an English grocery shop in the company of his small six-year-old grandson. Unable to express himself in English, he asks the small boy to help. The boy interprets for his grandfather, switching effortlessly between Gujarati and fluent Midlands English.

An Italian father addresses his son in Italian. The son listens carefully, then responds in English. The same son, forced to speak to another relative who does not understand English, is able to frame only a few elementary sentences in broken, strongly accented Italian.

Scenes such as these may serve as dramatic reminders of language shift, but they give no indication of the degree and extent to which it takes place. How, then, is the degree of language shift to be measured?

The short answer is by extensive language surveys, or by the inclusion of language-related questions in national censuses. Neither of these methods is ideal. Linguistic surveys are notoriously costly and difficult to carry out, and even the largest can only hope to draw upon information from a small sample of the population. The Australian Bureau of Statistics 1983 survey, for example, was based on a sample of two-thirds of 1 percent of the population aged over 15 years, the data being extrapolated to give figures for the entire nation.

Censuses, on the other hand, endeavour to question every single citizen of a nation. The number of questions which can be included on the census form is, however, severely limited, and those related to language are usually only a fraction of those included in dedicated linguistic surveys. In consequence, censuses provide us with a glimpse of only one or two of the most visible aspects of language use, without furnishing much of the supplementary information necessary to put them into perspective. As the authors of a major linguistic survey of minority languages in England point out (Linguistic Minorities Project 1985), 'language censuses of adult populations at best provide a broad social and geographic picture of reported language and literacy skills, without telling us anything about the social processes involved in patterns of language use in different spheres of activity'.

A word needs to be said about some of the inherent limitations of linguistic data derived from both censuses and surveys. Although they differ in form and purpose, both suffer from the same fundamental drawback of relying upon subjective assessment of linguistic ability and patterns of language use. As has been mentioned in a previous study of language choice (Buda 1991), respondents may not be fully conscious of their own language usage patterns, or may wish to portray them in a socially or culturally favourable light. Very often the respondent's assessment of his or her own language ability and usage represents more of what he or she would wish them to be, and less of what they really are. Subjectivity is also an unavoidable concomitant of surveys in which the researcher makes an assessment of the language ability of the subject.

A further problem of data-gathering methods based upon questionnaires is that of the way in which questions are framed. A restrictive form of question requiring only one answer might result in a loss of valuable data, whilst a more open-ended form of question might result in a flood of data which proves almost impossible to analyze.

To give an example, the Canadian Census of 1981 asked respondents to indicate only one mother tongue and home language; in spite of these explicit instructions, 597,980 persons (2.5% of the population) insisted on reporting more than one mother tongue and 535,735 persons (2.2% of the population) reported more than one home language. The 1986 Census introduced new guidelines under which individuals could report more than one mother tongue if they had learned them at the same time and had spoken one as frequently as the other when they were children. Similarly, respondents could indicate more than one home language if they were now speaking them equally often at home.

The result of this change in question was that the number of multiple responses given in the 1986 Census was significantly higher than in the 1981 Census. In 1986, 954,940 persons or 3.8% of the population reported a multiple response to the mother tongue question, while 1,159,675 or 4.6% of the population indicated more than one home language (Statistics Canada 1989).

It also needs to be pointed out that English-language questionnaires which ask the respondent to evaluate his or her fluency in the English language are highly unlikely to evince meaningful responses from those whose English ability is insufficient for them to understand the question.

Even where due care has been taken in the framing of questions in censuses and surveys, it is sometimes difficult to draw meaningful conclusions from the data obtained. To obtain some idea of language shift within a particular ethnic group, it is necessary to define first both the group and the languages it uses. This is often not as easy as it may seem. Multiple countries of origin, coupled with increasing intermarriage in successive generations, can blur the outlines of an ethnic group.

Defining a language can also prove extremely difficult, as can the definition of exactly what constitutes 'a speaker' of a particular language. For example, a rudimentary vocabulary and a few memorized phrases or expressions can often be sufficient to enable basic communication. On the other hand, a profound knowledge of a language does not necessarily guarantee oral fluency. Where is the line to be drawn between familiarity and fluency, and who is to draw it?

For all their inadequacies, language surveys and censuses remain the only methods we have to measure language use among large sections of the population. Bearing in mind some of the reservations mentioned above, we now turn to an examination of the relevant data for Australia and Canada.


The ethnic origins of the Australian population are complex. The British Isles (England, Scotland, Wales, the Isle of Man) and Ireland account for nearly 75% of a total population of 16.3 million (as at 26 January 1988). Nearly 20% of Australians are of continental European origin, and the remaining 5% of Asian, African, American, and Pacific origin. The indigenous Aboriginal population numbers 163,000, exactly one percent of the total population.

Viewed by country of origin, and leaving aside the dominant Anglo-Celtic group, the largest ethnic community is the German (618,000), followed by the Italian (605,250), Greek (323,110), Dutch (235,330), and Chinese (196,310).

These figures do not, of course, reflect the linguistic status of each group. The Australian Language Survey of 1983, for instance, showed that only 166,500 respondents gave German as their mother tongue, and only 79,900 reported using the language at home, thus placing German in third place behind Italian (440,800 and 360,800 respectively) and Greek (227,900 and 204,000).

The figures for ethnic origins do, however, provide a useful starting point for any attempt to measure language shift within each of the ethnic communities.

1976 Census Data

The first Australian census to elicit data on the regular use of community languages was that of 1976.

The census found that the most widely used community language was Italian, although figures for the individual States of the nation showed that German was the most widely used language in Tasmania and the Australian Capital Territory.

In Victoria, New South Wales, and South Australia, Greek was the second most widely used language, while in Queensland and South Australia second place was taken by German.

The census also revealed a significant difference in the pattern of language use in the nation's two major cities. The figures for Melbourne showed a predominance of the major ethnic languages (Italian, Greek, German, etc.), whilst in Sydney the most important languages were Arabic, Chinese, and Portuguese (ranked 8th, 12th, and 18th nationwide)

Cross-tabulation of the data obtained from questions about (i) birthplace and (ii) languages regularly used provided a rough but valuable indication of language shift patterns among first generation members of ethnic minority groups. A similar cross-tabulation of responses to questions about (i) birthplace of parents and (ii) languages regularly used gave an insight into the scope of language shift in the second generation, and further indicated the significant influence of exogamy upon language maintenance (see Table 1).

Table 1 Percentage Shift to English — Australia
Birthplace of parents
First Generation
Second Generation
(Two parents) a
Second Generation
(One parent) b
a. Australian-born, with two parents born in one or two selected countries
b. Australian-born, with one parent born in a selected country and the other parent Australian-born, British-born or Irish-born.

In "Language Policy and Community Language", M. G. Clyne (1988) draws the following conclusions from the 1976 Census data:

Some ethnic groups maintain their language better than others, according to a rank ordering (Greeks, Italians, Yugoslav groups, Poles, Germans, Maltese and Dutch), and this order does not differ markedly between the first and second generations or in different States. There is a very substantial shift towards English monolingualism between the first and second generations, and the most substantial shift is where one parent is Australian-born, British-born or Irish-born. Some States (for example, South Australia and Victoria) generally maintain community languages better than others (for example, Queensland and Western Australia). Language shift tends to increase with age in the second generation and to decrease with age in the first generation. This is related to life-cycle factors, including reversion to the first language among some elderly migrants.

Even so, the results do contain a few interesting anomalies. According to the data for language shift to English in the first generation, Tasmania exhibited a relatively high shift for those born in Italy and Yugoslavia, but a relatively low shift for those born in Poland. Whereas the average nationwide shift from German to English among second-generation members with two ethnic parents was 62.3%, Queensland showed an extraordinarily high shift of 84.1%, and the Australian Capital Territory a correspondingly low shift of 42.8%. No such anomaly was evident among second-generation members with only one ethnic parent, Queensland exhibiting a shift of 97.2%, and the Australian Capital Territory a shift of 95.6%, both figures very close to the national average of 96.2%.

Australian Bureau of Statistics 1983 Survey

Entrusted with the task of considering 'The Development and Implementation of a Co-ordinated Language Policy for Australia', the Senate Standing Committee on Education and the Arts requested the Australian Bureau of Statistics to carry out a nationwide language survey. The preliminary results of this survey were released in December 1983, and were the cause of immediate controversy.

Prior to the Australian Bureau of Statistics survey of 1983, most estimates of English language ability among migrants had been based upon the findings of the 1973 Immigration Survey. In 1978, for example, the Ethnic Affairs Commission (N.S.W.) reported that approximately 300,000 migrants in Australia, and 100,000 in New South Wales, were non-fluent in English or spoke 'fair to poor' English. In the same year the Galbally Report of the Review of Post-Arrival Programs and Services for Migrants estimated that 300,000 people needed English language instruction, and 400,000 permanent residents of Australia had 'little or no fluency' in English. Four years later, in 1982, the Australian Department of Immigration and Ethnic Affairs estimated that 'some 525, 000 (people) would appear to have less than survival skills in the four areas of English language competence'(Crowley 1984).

The Australian Bureau of Statistics survey of 1983, on the other hand, reversed this trend of creeping estimates by concluding that nationwide 65,800 people did not use, and 47,100 did not understand English. Many people involved with the provision of English language services for migrants were concerned that these figures might be used as a basis for a cutback in such services, and the figures of the survey were subjected to intense scrutiny. The survey has stood up relatively well to such scrutiny, and to a comparison with other census and survey results, and provides us with a useful corpus of data for estimating the magnitude of language shift among migrant communities in Australia.

Some of the survey's major findings having direct relevance to language maintenance were:

Of people whose first language was not English, 1,395,200 (74%) still speak the language at home; 1,734,100 (92%) still speak the language socially; 1,216,100 (64%) still read the language; 1,205,700 (64%) still write the language; 485,900 (25.5%) speak but do not read or write the language; 23,500 (1.2%) neither use nor understand the language (Crowley 1984).

Turning to specific major ethnic language groups (see Table 2), if we take the ratio of first language (the mother tongue) to language still spoken at home (the home language) to be an indication of language maintenance, and tabulate the derived percentages in order of descending magnitude, it becomes clear that neither size of ethnic group nor number of speakers claiming that language as a mother tongue has a direct correlation to degree of language maintenance.

Table 2 Language Maintenance Rates — Australia
Mother Tongue
Home Language

This conclusion contradicts the general belief that retention of language is related to the opportunities for its use, and consequently to the opportunities to meet and converse with other speakers of that language. Larger groups should, theoretically, exhibit a higher degree of language maintenance. Whilst not invalidating this theory, the Australian figures suggest that size is not the only factor involved in language maintenance.

For example, the largest minority language groups in Australia are the Italian, Greek, German, and Dutch, yet the German and Dutch groups have the lowest language maintenance figures of all (48% and 41% respectively). Conversely, the Greek group has the highest (90%), whilst the largest group of all, the Italian, comes sixth (82%). The average rate of language maintenance for the fourteen largest groups is 73%, a figure which emphasizes the rate of language loss among the Dutch and German communities.

The above figures refer to the use of community languages as an oral medium of communication. Taking into account the added dimension of degree of literacy in both English and the heritage language, the survey further found that:

English literacy is highest for Dutch, German and French speakers, and lowest for speakers of Vietnamese, Arabic and other Yugoslav languages (i.e. Slovenian and Macedonian); a correlation exists between high rates of language maintenance and a high degree of literacy in the heritage language; literacy in their first language is lowest for Maltese, Dutch and Polish speakers; literacy is particularly low for languages used principally in informal situations (e.g. Maltese), or ones with large numbers of dialect speakers (for example, Italian) (Crowley 1984; Jupp 1988).

The Ethnic Affairs Commission of NSW drew the following conclusions from the findings of the 1983 Language Survey:

The figures indicate that 23,500 people in Australia who first spoke a language other than English no longer speak or understand that language. Further, 485,900 people speak but do not read or write their first language (that is, either they have never acquired literacy skills in their own language, or they may have once been literate but have since lost those skills). If one accepts as the Ethnic Affairs Commission does, that a multilingual population represents a major national resource, and that people should be encouraged and assisted to maintain and/or develop their language skills, then these figures do indicate the need for the development of programs in language maintenance and development, and also which groups are most likely to need or want such assistance (Crowley 1984).


Before examining the ethnic origins of the Canadian population, it should be noted that prior to the 1981 Census, respondents were asked to state only their paternal ancestry. In cases where respondents insisted on stating multiple ethnic origins, only one origin was accepted, resulting in a single ethnic origin per respondent. This restriction was removed in the 1981 Census, and one write-in option was provided. The 1986 Census allowed respondents to write in up to three ethnic origins not included in the mark boxes (Statistics Canada 1989).

The 1986 question on ethnic origins was altered slightly from that posed in the 1981 Census. In 1981, respondents were asked, "To which ethnic or cultural group did you or your ancestors belong on first coming to this continent?" The phrase "on first coming to this continent" was deleted from the 1986 question. The 1986 ethnic origin question was: "To which ethnic or cultural group(s) do you or did your ancestors belong?"(Statistics Canada 1989)

Bearing in mind that in 1986 the total population of Canada was 25,354,064 (by 1988 it had grown to 25,923,300 ), we can now turn to an analysis of the nation's ethnic composition.

Of the 18 million people giving a single response to the 1986 census question on ethnic origins, 6,332,725 (35.1%) stated they were of British origin, and 6,093,160 (33.8%) said they were of French origin.

The other major single-response ethnic groups were: German (896,720), Italian (709,590), Ukranian (420,210), Chinese (360,320), Dutch (351,765), South Asian (266,800), Jewish (245,855), Polish (222,260), Portuguese (199,595), and Scandinavian (171,715).

Over 13 million respondents gave multiple answers to the same question, and a cross-tabulation of single and multiple answers affords an insight into the homogeneity of each ethnic group.

Groups with a significantly low ratio of multiple to single origin answers were the Korean (0.07), Bangladeshi (0.13), Cambodian (0.14), Filipino (0.15), Chinese (0.15), Laotian (0.16), Iranian (0.18), East Indian (0.18), and Vietnamese (0.19). The European group with the lowest ratio was the Portuguese (0.19), followed by the Greek (0.23), Croatian (0.26), French (0.33), and Italian (0.42).

The homogeneity of most of the above groups can be attributed to the relatively short time they have been in Canada, insufficient to allow for the birth and maturation of a second generation, and hence for a significant level of intermarriage with members of other ethnic groups.

The European groups listed above are, however, relatively mature communities (the French particularly so), and the low ratio of multiple answers points to a resistance to exogamy.

Conversely, groups with a high ratio of multiple to single origin answers were the Swedish (3.70), Norwegian (2.96), Icelandic (2.71), Russian (2.24), Swiss (2.15), Austrian (2.00), and Danish (1.98). It can be assumed from these figures that these groups have a proclivity to marriage outside their own community. The prominence of Scandinavian groups is highly significant. To place these figures in perspective, it should be noted that the non-European group with the highest ratio was the Malay (1.92), followed by the Indonesian (1.79) and Mexican (1.71).

In addition to allowing multiple responses to the question on ethnic origins, the 1986 Census also allowed multiple responses to questions on mother tongue and home language.

Although the census did not ask questions about language maintenance as such, a cross-tabulation of both the single and multiple responses to questions on ethnic origins, mother tongue, and home language can provide some important pointers to the extent to which heritage tongues are being replaced by either English or French, and to the influence of exogamy on language maintenance.

1986 Census

Looking first at the responses to the questions on language, and leaving aside the dominant official languages (English and French), the language most frequently reported as a mother tongue was Italian (455,820), followed by German (438,675), Chinese (266,560), Ukranian (208,410), Portuguese (153,990), Dutch (123,665), Polish (123,120), Greek (110,350), Hungarian (69,000), and Punjabi (63,640).

The most common home languages were Italian (271,835), Chinese (230,480), German (112,550), Portuguese (105,420), Greek (72,550), Spanish (55,760), Polish (55,150), Punjabi (47,865), Ukranian (46,150), and Vietnamese (40,345).

As with the corresponding Australian data, these figures do not, in themselves, give an indication of the degree of linguistic retention within each ethnic group.

Once again we must turn to the census data for mother tongue and home language to ascertain the percentage of ethnic minority members still using their first language at home, and hence the level of heritage language maintenance (see Table 3–1).[1]

Table 3–1 Maintenance Rates of Major Languages — Canada
Mother Tongue
Home Language

The above figures are derived from adjusted language data. In other words, the figures for mother tongue and home language include both single and adjusted multiple responses. It is not clear how many of the respondents included in each of these figures acquired, or still use, more than one ethnic language. What is more, it is impossible to ascertain whether there is any difference in fluency or frequency of use among the several languages given in multiple responses.

Consequently, the above figures should not be interpreted as absolute indications of the degree of language shift within each ethnic community. They do, however, suggest that striking similarities exist between language maintenance patterns in Australia and Canada. Vietnamese, Chinese, and to some extent Greek show relatively high rates of maintenance in both countries, while German and Dutch rates are relatively low. Once again, there would seem to little or no relation between the size of the language group and the rate of maintenance. A confirmation of this surprising result is obtained by comparing the average maintenance rates of the twenty largest language groups (Table 3–1) with that of the twenty smallest (Table 3–2). The resulting figures are 52% and 48% respectively.

Table 3–2 Maintenance Rates of Minor Languages — Canada
Mother Tongue
Home Language

An examination of statistical profiles of ethnic groups derived from 1986 Census data shows that the figures for mother tongue and home language are not directly related. For example, the number of respondents within each ethnic group answering the question on mother tongue was not necessarily the same as the number of respondents in the same group answering the question on home language. Intermarriage, for example, can result in a person using an ethnic home language which is not his or her mother tongue. Likewise, it is conceivable that a member of an ethnic group is unable to use his or her mother tongue at home, but is able to use it on a daily basis at his or her place of work.

The ethnic group profile statistics do, however, enable us to compare the sizes of ethnic groups with the total number of group members claiming the relevant ethnic language as a mother tongue.

The ethnic groups with the highest percentage of total mother tongue responses were the Vietnamese (87%), Cambodian (85%), Laotian (84%), Chilean (84%), and Punjabi (83%). Those with the lowest were the German (17%), Belgian (10%), Icelandic (7%), Swedish (6%), and Norwegian (5%).

These figures show a marked correspondence to the figures given earlier for ethnic homogeneity, and confirm that exogamy plays a significant role in loss of heritage tongues. There were, however, a number of exceptions to this correlation between mother tongue maintenance and ethnic homogeneity. The Thai group showed high rates of both language maintenance and exogamy, as did the Lithuanian, Spanish, and Punjabi groups. Conversely, groups combining high homogeneity with relatively low rates of language maintenance were the Slovenian, Tagalog-speaking Filipino, and Slovak.


Having taken a look at the Australian and Canadian figures for language shift among ethnic communities, it may be useful to attempt a comparison of the two sets of data.

A comparison of ethnic and linguistic data derived at different times, in different countries, and in different ways, is, of course, fraught with dangers. It can be compared to a comparison of two snapshots taken of different people, in different locations, at different times of day, and with different cameras. Nevertheless, such a comparison does furnish some interesting correspondences from which a number of tentative conclusions may be drawn.

In both countries, recent immigrant communities show high rates of ethnic homogeneity and language maintenance—a not altogether surprising result. Notable examples are the South-east and East Asian groups in Australia and Canada, and the Central and South American groups in Canada.

Increased length of residence in the host country produces a language shift towards English.[2] This shift increases with the second generation, and is further accelerated by marriage outside the ethnic community.

The rate of shift is not, however, the same for all languages. Some groups seem to maintain their heritage language far better than others. Established groups with high rates of mother tongue maintenance in both Australia and Canada were the Greek, Macedonian/Slovenian, and Croatian. In Canada the Lithuanian and Armenian groups also showed high rates of maintenance.

At the other extreme, Dutch figured prominently at or near the bottom of the maintenance tables for both Australia and Canada. In Canada, Swedish, Danish, and Norwegian rates of maintenance were extremely low.

A comparison of the average maintenance rates of 12 selected languages gives a figure of 74% for Australia and 57% for Canada. Languages exhibiting a statistically significant exception to the tendency for retention rates to be higher in Australia than in Canada were Vietnamese, Chinese, and Serbian (see Table 4).

Table 4 Comparison of Language Maintenance Rates — Australia and Canada

No single factor can account for the differing individual language maintenance figures, or for the apparent difference in maintenance rates between Australia and Canada.

Cross-reference to figures for size of community and geographic distribution and concentration does not provide any conclusive correlation. Whereas the German community in Australia is numerically the largest ethnic group, it has a language maintenance rate of 48%, the only group with a lower rate being the Dutch—the fourth largest group.

The same disparity can be seen in the Canadian data, with some relatively large ethnic groups such as the German, Ukranian, and Dutch having low maintenance rates (26%, 22%, and 12% respectively), whilst some relatively small established communities such as the Lithuanian and Armenian have extremely high rates (81% and 78%).

Data for both Australia and Canada suggest strongly that ethnic homogeneity and language maintenance are related, though the influence of other factors does produce exceptions to this tendency.

Other factors which could conceivably have an effect on language maintenance are the cultural or religious vitality of the community and the opportunities available for use of the language.

Use of a specific language could be tied to a religion or cultural value system which exhibits more vitality than the language itself. The Punjabi-speaking community in Canada, for example, exhibits significantly higher rates of language maintenance than any of the other major Indian languages, and the highest rate (79%) of any Canadian ethnic language for multiple-response maintenance of the home language. One explanation for this could be the influence of the Sikh religion.

Opportunities for language use are difficult to define and quantify. The home provides the first and most important environment for use of a mother tongue. The attitude of minority group parents towards their heritage language will determine how effectively they transmit that language to their children. The first years at school are a particularly critical time for language maintenance, with daily exposure to the majority language having a powerful restraining effect upon further development of mother tongue abilities. Heritage language education during the formative pre-school years is a vital first step if that language is to preserved. Often this early exposure to the mother tongue is insufficient in itself, and needs to be reinforced by additional semi-formal instruction. Not many minority groups enjoy the luxury of their own schools, but many seek to compensate by organizing weekend classes conducted by qualified volunteers. These classes concentrate on heritage culture and language, and provide an important means of carrying over a knowledge of both into the teenage years.

Once the children of ethnic communities leave school, the opportunities for heritage language use are dramatically reduced, particularly if the children also leave home at this time. It is at this stage that institutionalized support of language use plays an important role. Social clubs and community centres, minority language newspapers and periodicals, radio and television programmes, lending libraries and video rental facilities—all these help to reinforce the heritage language, as do religious and social activities associated with local churches, temples, or mosques. Even where members of a particular ethnic community belong to a mainstream Christian denomination, they may often prefer to set up their own churches where they can worship and commune in their heritage tongue.

The provision of institutionalized support such as this requires money. Although limited amounts of government funding exist in both Australia and Canada, most of the money comes from donations and fund-raising activities such as lotteries, concerts, and bazaars. How much money a specific community is able to raise toward the provision of such facilities is dependent on not only that community's awareness of the need for such linguistic and cultural reinforcement, but also on that community's social and economic resources. The authors of the Linguistic Minorities Project (1985) see these factors in a wider context:

Factors affecting language maintenance may be fundamental external forces influencing the economic, political and social resources of a linguistic minority. There may also be factors emerging from within the minority; and, in a situation of contact between cultures, this 'internal' response is often a reaction to outside pressures. Similarly, the process of 'language shift' is a reflection of power relations in a society producing a redistribution of social and linguistic resources from generation to generation.

It is relatively easy to identify many of the factors contributing to language maintenance; it is far more difficult to determine which of them play a role in the maintenance of specific ethnic languages, and to what degree. To what combination of factors should we attribute the resilience of Greek in both Australia and Canada (maintenance rates of 90% and 66% respectively), or the surprisingly low maintenance rates of Dutch (41% and 12%)? Likewise, how do we explain the fact that the differences in comparative maintenance rates between Australia and Canada are larger for some languages than for others? Of the twelve selected languages in Table 4, for instance, Arabic has an above-average maintenance rate disparity of 87% in Australia and 54% in Canada.[3] On the other hand, the respective rates for Chinese are 84% and 86%, and for Vietnamese 89% and 97%, both well above the respective national averages, and both in contrast to the tendency for maintenance rates to be lower in Canada than in Australia.

Indeed, how to explain the overall differences in maintenance rates between Australia and Canada? Any study seeking to address this question would have to take into account the social, cultural and economic characteristics of the two nations as a whole—particularly their attitudes and policies towards immigrants and ethnic minorities. This is well beyond the scope of the present study, and this question will have to remain unanswered for the time being.

This study has attempted to identify some of the most prominent features of language shift in Australia and Canada without commenting on either the inevitability of the phenomenon, or the advisability of concerted government efforts to preserve ethnic languages and cultures. It is hoped to deal with these issues in a future study.


Buda, J. K. 1991. Language Choice. Otsuma Review 24.

Clyne, M. G. 1988. "Language Policy and Community Language." In The Australian People: an Encyclopedia of the Nation, Its People and Their Origins, ed. J. Jupp. North Ryde NSW: Angus and Robertson.

Crowley, Desmond. 1984. Language Survey: the Ethnic Affairs Commission's Comment. Occasional Papers 2.

Linguistic Minorities Project. 1985. The Other Languages of England. London: Routledge & Kegan Paul.

Statistics Canada. 1989. Profile of Ethnic Groups. Ottawa: Statistics Canada.


1  The present study is concerned with language maintenance among immigrant communities in Australia and Canada. For this reason data on language maintenance among Aboriginal peoples (Inuit, Métis, and North American Indian) have been excluded.

2  There is also some degree of language shift towards French among ethnic communities resident in Quebec. According to the 1986 Census, the largest ethnic community in this province was the Italian, numbering 163,880. 33,900 ethnic Italians gave French as their mother tongue, and 21,820 gave French as their home language.

3  The size of the Arabic-speaking community in Australia is approximately 104,000, and of that in Canada approximately 150,000.

Reprinted from Otsuma Women's University Annual Report: Humanities and Social Sciences, Vol. XXIV, 1992