In the past, marriage was close to universal in most Asian countries. The proportion of women aged 45-49 remaining unmarried exceeded 2% only in Sri Lanka, Myanmar and the Philippines (Smith, 1980). But this is no longer the case. Over the past two decades some dramatic changes have taken place in Southeast and East Asia: non-marriage for women is becoming much more common, and in many of the big cities. it is even more common than it is in Western countries, notwithstanding the sharp declines in marriage prevalence in Western countries in recent times. The main reason why observers have been slow to recognize this fact has been that the index of non-marriage normally used (percent of women remaining unmarried near the end of their reproductive period) reflects the high marriage probabilities in the past, since most marriage activity for a cohort reaching age 45-49 occurred a quarter century earlier. Such a measure will fail to reflect major changes in non-marriage from older to younger cohorts until those younger cohorts reach ages in their 40s. But it is possible to detect such changes ten or fifteen years earlier, by studying trends in percentages remaining unmarried in their early 30s, a statistic which bears a fairly predictable relationship with the percentage of the same cohort remaining unmarried in their 40s. This close relationship can be inferred from Coale's discovery that there are precisely defined age patterns of nuptiality (Coale 1971; Coale and McNeil. 1972). Coale did not apply his analysis to the issue in question, because he was interested in determining regularities in nuptiality patterns across populations differing in certain characteristics, including their proportions ultimately marrying, but not in studying how those patterns changed when a given population underwent large changes in the proportion ultimately marrying. But the relationship can easily be studied empirically. For example, in Thailand. the ratio between the proportion single among females aged 45-49 and the proportion single in the same cohort ten years earlier averaged 0.75 in 1970 and 1980. If this ratio continued to hold in future. then the proportion single among women aged 35-39 in 1990 would result in higher proportions single for this cohort 10 years later than had been the case for cohorts reaching age 4549 in 1980 and 1990. Calculations for the other countries yield similar results. The paper will deal mainly with non-marriage for women, though the trends in non-marriage for men will also be dealt with more briefly. The focus will be on the two main cultural blocs of East and South-East Asia: the Confucian world of sinic cultures, and the Malay world. In the first, the family tended to be more patriarchal, with women largely restricted to the home. Most Chinese families were extended joint or stem) at some point in their life cycles (Freedman (ed., 1970; Wong, 1978; Huang, 1992). In the Malay world, there was more emphasis on the elementary nuclear family, the kinship system was bilateral, and women had greater autonomy.
