Michael Clyne and Sandra Kipp

There is a continuing significant shift to English spoken in the home among Australia’s established community language groups. There are also success stories in language maintenance. Factors influencing language use include the distribution of speakers, the age profile of the community, intermarriage patterns and cultural distance from Anglo-Australians. Australia-wide, the shift rates to English spoken at home range between three per cent from Macedonian and 62 per cent from Dutch in the first generation, and 15 per cent from Macedonian and 95 per cent from Dutch in the second generation.

INTRODUCTION

Analysis of census statistics in Australia² and overseas³ has pointed to substantial variation in the extent of maintenance of community languages. The continuing use of the same language question in the past three Australian censuses makes the study of trends possible.

In an article in the previous issue of People and Place⁴ we analysed some main features of the language data from the 1996 Census. We suggested that the rises and falls in the number of people using community languages in the home may be due at least in part to fluctuations in the patterns of language maintenance and language shift (from the community language to English). We are here able to confirm the significance of this factor for several of the languages. The language statistics we have been able to calculate from the 1996 Census are far more extensive than those from previous Australian censuses in that they cover more ethnolinguistic groups (for example Japanese, Korean, Macedonian) and provide us with much more detailed information about the second generation. Through comparisons with 1991 data we are able to verify a continuing variation in language shift along a cultural-regional continuum. We are also able to identify differences between the first and second generations, interstate differences and variation in the second generation according to family marriage patterns. Variation between age groups, and according to period of residence, further facilitate the identification of factors promoting or impeding language shift.

LANGUAGE SHIFT IN THE FIRST GENERATION

The language question used in the last three Australian censuses has been: ‘Does the person speak a language other than English at home?’ In 1996 there was the added advice: ‘If more than one language, indicate the one that is spoken most often’. Language shift in the first generation is calculated as the percentage of persons born in a particular (non-English-speaking) country who now speak only English at home. The lack of a ‘language first spoken’ question necessitates the use of birthplace as a surrogate indicator of language background. This clearly renders the calculation of language shift figures for countries such as Viet Nam, ‘Yugoslavia’ or Egypt, impossible due to the diversity of languages spoken in the home country. For example, Viet Nam-born persons may speak either Vietnamese or Chinese as a first language, Egypt-born persons may speak Greek, Italian, French, Maltese or Arabic, and ‘Yugoslavia’-born persons may speak an even wider range of languages. Despite this difficulty, the remarkable reliability of the statistics is evidenced by the almost identical shifts among Austrians and Germans (within one per cent over two successive surveys, 1991 and 1996, covering the entire Australian population).
 

Table 1: Language shift in the first generation, 1991 and 1996
Birthplace	% shift 1991	% shift 1996
Austria Chile France Germany Greece Hong Kong Hungary Italy Japan Korea, Republic of  Lebanon FYROMb Malta Netherlands Other South America Poland PRCc Spain Taiwan Turkey	42.5 n.a.a 31.5 42.4 4.4 8.4 26.7 11.2 n.a. n.a. n.a. n.a. 31.0 57.0 n.a. 17.2 5.9 16.5 3.0 3.8	48.3 9.8 37.2 48.2 6.4 9.0 31.8 14.7 15.4 11.6 5.5 3.0 36.5 61.9 17.2 19.6 4.6 22.4 3.4 5.8
a n.a. stands for ‘not available’ b Former Yugoslav Republic of Macedonia c People’s Republic of China

The term language shift refers here to the proportion of a group born in a non-English-speaking country who, at the time of the census, spoke ‘English only’ at home. The rank ordering of the rate of language shift is, with very few exceptions, the same as in 1991. However, again with few exceptions, it has increased over the five-year period. The range extends from three per cent for those born in the ‘Former Yugoslav Republic of Macedonia’ (henceforth referred to as FYROM), a figure which was not available in 1991, to 61.9 per cent for those born in the Netherlands. Towards the Dutch end of the spectrum we have German-, Austrian-, French-, Maltese-, and Hungarian-born, while towards the Macedonian end we have those born in Taiwan, the People’s Republic of China (PRC), the Lebanon, Turkey, Greece, Hong Kong, Chile, and Korea. In the middle are those born in Italy, Japan, Spanish-speaking South America other than Chile, Poland and Spain.

As can be seen from Table 1, the clear exception to the overall rise in language shift is the PRC-born community, where the shift to ‘English only’ in the home has decreased from 5.9 per cent to 4.6 per cent, probably due to the large increase in migrants from that country over the past three years.⁵ We do not have 1991 data on language shift among the Koreans and the Japanese, which have experienced similar population increases over the same time. The only slight change in the rank ordering from 1991 to 1996 is due to the Spanish-born increasing their language shift by more than the Polish-born, again probably due to the larger new component in the Polish-Australian community (an eight per cent rise in the past five years compared with a 3.6 per cent rise among the Spanish-born). We shall discuss Greek and Macedonian Australians in the section on language shift for the second generation.

The above statistics underline the importance of cultural distance in determining language maintenance and shift.⁶ That is, those from predominantly Islamic or Eastern Orthodox cultures (Greek, Lebanese, Macedonian, Turkish) are more likely to maintain their languages at home than are other groups from Europe. Groups from northern, central and western Europe tend to shift to English the most. People from Asian countries, especially Chinese-speaking ones, tend to display a low or fairly low language shift. The shift for the PRC-born (of whom 39.7 per cent are Cantonese speaking and 36.4 per cent Mandarin speaking) is a little higher than for the Taiwan-born, who are predominantly Mandarin speakers (82.5 per cent), and that for the Hong Kong-born is three times as high as for the PRC-born. This may be at least partly due to the status of the fang yan, or variety, spoken (for example, Cantonese, Mandarin). Whereas Mandarin is China’s standard language, the variety with the high market place value⁷ and the language taught in Australian schools, most of the Hong Kong-born (86.5 per cent) use Cantonese as their fang yan, which has a lower perceived status outside of Hong Kong. In addition, their exposure to English has been much greater due to the British colonial status of their birthplace, the degree of westernisation and the low degree of national identity with Hong Kong.⁸

The shift to English in the home among people from Spanish-speaking South American countries continues to be much lower than among Spanish migrants. All Spanish-speaking groups have increased their shift since 1991.

REGIONAL DIFFERENCES IN LANGUAGE SHIFT

A comparison of state language shift figures (Table 2) and total population figures shows that languages are maintained best in the State in which they are best represented. For instance, Greek, Italian, Maltese, Turkish and Macedonian show the lowest language shift rate in Victoria, the State in which their proportion of the total population is the greatest. The same applies to speakers from Lebanon, Hong Kong and Korea in NSW, and from Germany and Poland in South Australia. In the majority of cases this pattern extends to the languages with the second highest population share in a particular State, which then also record their second lowest language shift rate in that State. Examples include German and Polish in Victoria, Maltese and Turkish in NSW and Italian and Greek in South Australia. Older established groups with the most uniform distribution patterns (Polish, German, Austrian, Hungarian and French) produce the most consistent language shift trends. The most consistently low shift across the States is recorded by the Taiwan-born (varying from 2.2 per cent to 8.1 per cent) as compared with those born in FYROM (two per cent to 16.7 per cent). This variation is probably due to the very high concentration of Macedonians and the relative recency of arrival of the Taiwan-born.
 

Table 2: Language shift by state, first generation, 1996
Birthplace	NSW	VIC	QLD	SA	WA	TAS	NT	ACT	Aust.
Austria Chile France Germany Greece Hong Kong Hungary Italy Japan Korea Lebanon Macedonia  Malta Netherlands Other Sth America Poland PRC Spain Taiwan Turkey	49.8 9.9 37.1 49.1 6.2 6.8 32 15.5 14.7 7.2 4.9 3.1 36.7 62.8 14.3 21.4 4.0 20 3.5 5.2	46.1 7.6 38.7 47.1 5.4 7.8 28.2 12.5 17.0 19.1 5.4 2.6 31.3 62.1 16.7 19.4 3.8 22.3 4.5 4.4	48.2 14.4 36.2 48.1 15.6 15.3 36.2 22.1 13.0 13.7 19.7 9.1 57.3 62.6 28.3 24.4 10.0 24.2 2.2 15.3	46.6 13.8 38.6 45.7 6.1 14.0 32.0 13.8 20.0 48.4 9.4 6.6 46.5 61.0 27.9 14.5 7.4 33.5 4.5 9.8	49.1 12.1 35.1 50.7 13.6 19.4 32.7 16.0 17.3 29.9 14.8 2.9 67.1 58.9 28.9 17.0 5.9 25.0 4.3 12.0	53.7 16.8 43.4 50.8 11.0 24.8 52.3 28.3 17.0 56.0 30.4 16.7 73.8 60.7 37.3 21.9 15.0 42.0 8.1 6.7	55.4 18.8 30.7 50.1 4.8 19.7 48.9 24.3 17.5 67.6 37.5 0.0 66.7 66.9 35.1 48.9 11.4 28.8 15.4 12.5	49.1 12.6 38.6 51.3 8.8 17.9 33.5 21.3 26.1 22.3 10.3 6.3 56.0 65.3 25.2 19.5 7.1 18.7 10.4 14.3	48.3 9.8 37.2 48.2 6.4 9.0 31.8 14.7 15.4 11.6 5.5 3.0 36.5 61.9 17.2 19.6 4.6 22.4 3.4 5.8

 

LANGUAGE SHIFT IN THE SECOND GENERATION

Language shift in the second generation is calculated as the percentage of persons born in Australia with one or both parents born in a particular country (identified with a particular language) who now speak only English at home. Figures are given for persons with both parents born in the same country (‘endogamous’), or with one parent born in that country and the other born elsewhere (‘exogamous’). Once again, birthplace has had to be used as a not particularly satisfactory surrogate for ‘language of background’.

A comparison between Tables 1 and 3 shows that, across all groups, there continues to be a high intergenerational shift and the near replication of the rank ordering of groups according to language shift established for the first generation. The main deviations are:

• The very substantial inter-generational shift among Italian-Australians, even in families where both parents are Italian-born. Second generation Italian-Australians from endogamous families show a shift three times as high as their parents’.

• The relatively low inter-generational shifts among Spanish speakers and French-Australians.

 
• High inter-generational shifts among Polish, Hungarian and Maltese Australians.

• Low shifts among those of Hong Kong, Japanese and Korean parentage. The anomalous situation where the second generation from endogamous families shift less than the first generation may be due to a number of factors, including the large student populations contributed by these birthplaces in the first generation (usually accompanied by mixed households), and a larger incidence of exogamy in the first generation of early vintage. This is particularly likely to hold for situations where the migration was primarily of single young adults and possibilities for endogamy may have been limited. In the case of Hong Kong families, the low second-generation shift is also probably motivated by continuing close business and social contacts entailing frequent travel to the former homeland.

• Relatively low second-generation shifts among second generation Australians of Taiwanese and Turkish parentage.

Table 3 indicates the importance of exogamy as a factor in language shift. In all the languages under consideration, the second-generation shift is considerably lower for the children of endogamous marriages. The difference is particularly marked for those ‘newer’ languages with very low shifts in the first generation (Hong Kong, Japan, Korea, Taiwan) and for those with relatively low rates of exogamy (Turkey, Lebanon). Some 88.7 per cent of second-generation Turkish Australians are the product of endogamous marriages, as are 73 per cent of second-generation Lebanese Australians, compared with 12.7 per cent of those with Austrian parentage and only eight per cent of those with French parentage.
 

Table 4: Language shift in the second generation (aggregated), 1991 and 1996
Parents’ birthplace	Language shift, 1991	Language shift, 1996
Germany Greece Hong Kong Italy Malta Netherlands Poland PRC	88.7 21.8 40.0 49.8 78.5 95.0 74.4 45.5	89.7 28.0 35.7 57.9 82.1 95.0 75.7 37.4

Comparisons with 1991 data are limited by the sparsity of cross-tabulations available from the earlier statistics. However, Table 4 indicates that the most dramatic departure from the patterns of the previous censuses has been the substantially increased language shift from Greek. Still the best maintained of the ‘older’ community languages, Greek now shows a shift rate of 6.4 per cent in the first generation and 28 per cent in the second generation (up from 22 per cent in 1991). This would seem to indicate that home language shift, while it may be slowed down by cultural values and successful language maintenance efforts, will occur eventually. Indeed, much of the above variation could be explained in terms of the relative age profiles of the second generation groups involved. For example, 49 per cent of Australians of Italian background are aged between 25 and 45. Not only are they of the age to have established their own households, but the third generation is also well established. The second generation of Taiwan-born parents, on the other hand, is still very young, with 87.1 per cent under the age of 15 and still within the family setting. Another traditionally well maintained language that has experienced a significant rise in language shift between generations is Turkish (5.8 per cent shift first generation, 16.1 per cent second generation). On the other hand, people from the FYROM, on whom we have statistics for the first time, still show a very consistently low language shift in both generations.

Although the majority of languages (except Dutch, where the rate was already 95 per cent in 1991) have seen an increase in the second generationshift between 1991 and 1996, the reverse is true for those whose parents were born in Hong Kong or in the People’s Republic of China. This could be explained by an increasing number of the second generation born of endogamous marriages as these communities have become established.

LANGUAGE SHIFT AND AGE

The 1986 Census⁹ showed that it was generally the oldest age group in the first generation that maintained the community language most and the oldest age group in the second generation that maintained it least.¹⁰ The former still applies in some communities, notably Italian, Maltese, Spanish and Turkish Australians, while in others (for example, Greek, Korean, Polish, Taiwan, Latin American) the 55-64 year-olds shift least, and in still others it is school and preschool children (for example, German-speaking, Dutch, Macedonian, Turkish). (See Table 5.) The first-generation age groups with the highest shift are the 25-34 year-olds (French, Greek, Latin American, Macedonian, Maltese, Turkish), but in some of the older, less retentive language groups a bigger shift occurred with the groups now 35-44 (Austrian, Dutch, Italian) or even 55-64 (German and Polish).
 

Table 5: Language shift by age (first generation), 1996
Birthplace	0-4	5-14	15-24	25-34	35-44	45-54	55-64	65+
Austria Chile France Germany Greece Hong Kong Hungary Italy Japan Korea  Lebanon FYROM Malta Netherlands Other Sth America Poland PRC Spain Taiwan Turkey	22.3 15.2 24.5 21.5 11.6 19.8 16.7 16.4 13.1 41.8 3.9 1.5 27.4 39.0 32.4 3.9 2.1 21.3 14.4 2.6	24.5 9.8 30.3 31.5 10.8 9.1 6.2 22.8 13.4 41.9 3.6 1.0 39.7 54.4 16.7 4.4 1.7 29.7 3.6 3.2	33.4 8.1 31.9 37.7 9.9 6.5 16.3 19.2 13.7 9.1 3.3 2.9 40.2 49.6 13.9 11.1 1.9 23.9 3.1 4.6	53.1 17.5 41.0 50.5 14.4 10.6 23.6 29.3 13.8 7.2 6.5 6.9 51.4 61.5 23.9 14.2 2.8 36.1 5.7 7.4	63.0 9.1 38.2 59.9 13.6 7.1 36.5 30.6 13.6 4.1 5.6 3.7 48.0 77.3 17.9 15.1 3.2 36.6 2.4 5.8	61.9 6.4 41.3 62.9 6.2 8.4 43.5 20.8 18.5 2.4 6.3 2.1 40.8 76.4 14.5 22.7 8.1 16.2 2.1 4.9	43.4 5.4 30.7 38.1 2.9 18.6 37.5 8.3 22.1 3.2 4.3 1.1 28.3 58.8 11.3 27.6 8.2 12.5 1.7 4.6	33.6 6.8 34.7 31.7 3.4 20.9 23.5 6.5 32.1 5.2 6.2 1.3 19.2 40.5 16.7 22.6 6.8 12.0 3.1 7.2

In the second generation (shown in Table 6), some groups, such as German and Spanish speakers, and those of Italian, Lebanese, Maltese, Hong Kong, Macedonian or Polish background, continue to show the earlier tendency of shifting most in the oldest age group or the 55-64 age group, where there are very few (or none at all) older generation family members in the home with whom to speak. However, there is now sufficient variation to warrant detailed study. The high shift among the 45-54 year-olds may reflect the assimilation of many families who came to Australia prior to World War II.

The lowest shift is generally recorded in preschool and/or school age children. However, children of the Austrian-, French-, German-, Hungarian- and Dutch-born tend to shift to English as their home language either on entering school or during their school years. There are two periods of increased shift for second generation Australians of Hong Kong, Japanese or Polish background — 5-14 (school age) or 15-24. This latter group is likely to be either employed or involved in tertiary study, and may also have moved out of the family home. For those of Korean or Taiwanese background, the 15-24 age group generally constitutes the time of shift.

Back to People and Place Home Page