PRE-TRANSITION FERTILITY IN CHINA:
OLD WINE IN NEW BOTTLES
Cameron D. Campbell
Sociology, UCLA
Wang Feng
Sociology, UCI
fwang@uci.edu
James Z. Lee
History, Caltech
jzl@its.caltech.edu
History and Sociology, University of Michigan
Pre-transition Fertility in China: Old Wine in New Bottles
Cameron D. Campbell
Wang Feng
James Z. Lee
Almost
two decades ago, in this journal and elsewhere, Arthur Wolf and Ansley Coale
engaged in a debate over the level of marital fertility in pre-revolutionary
China (Coale 1984, 1985; Wolf 1984, 1985).[1] Coale and his colleagues, George
Barclay, James Trussell, and Michael Stoto, after reanalyzing survey data
collected from 40,000 farm families in 119 widely dispersed localities in China
in 1929-31, concluded not only that marital fertility was lower in China than
in historical Europe, but that such a low level “would be expected by
demographers only in populations where some combination of contraception and
abortion is practiced (Barclay et al. 1976, 625).” Wolf, based largely on his own studies
of early twentieth century populations from Taiwan and his 1980-81 survey of
580 elderly women from seven other provinces, challenged their results and
their explanation.
During the last decade, in this journal and elsewhere, we and others have expanded our understanding of Chinese population behavior.[2] While it is still premature to produce reliable estimates of population size and demographic rates for China as a whole before 1950, contemporary national surveys and historical micro studies reveal consistent patterns of behavior over long periods of time for a variety of Chinese populations. In Lee and Wang (1999a and 1999b), we survey and summarize the demographic evidence from these historical micro-studies and compare their behavior with national trends for contemporary China.[3] We identify four distinctive aspects of Chinese demographic behavior—lopsided mortality and nuptiality, especially by gender, low marital fertility, and high rates of fictive kinship and adoption—that persist today, that differ from Western patterns, and that temper the classic Malthusian understanding of comparative demographic behavior in general and China in particular. [4] We demonstrate how these behaviors interacted historically to form a consistent demographic system that was deeply embedded in the collective nature of Chinese social organization. We construct, in other words, a stylized model of a Chinese demographic system to contrast with the ideal European demographic system first proposed by Malthus and elaborated by others, notably Hajnal (1965, 1982), Laslett (1983, 1988), Macfarlane (1978, 1986, 1987, 1997), and Wrigley et al. (1981, 1997).[5]
While
Wolf accepts our general characterization of Chinese population behavior, he
dismisses our evidence and disputes our explanation of low marital fertility
(2001, 134).[6] As in his earlier exchange with
Coale, he argues that that the quality of our data is
poor, that our estimates of fertility are low, that our results are
unrepresentative of China as a whole, and that “the remaining difference
is due to positive rather than preventive checks (1984, 445).” [7] Wolf rejects the possibility that
Chinese couples in the past deliberately sought to control their fertility, and
argues, “far from limiting the number of sons reared, Chinese families
made every effort to maximize the number” (2001, 134). In so doing he largely repeats his 1984
assertions and arguments, while ignoring Coale's 1984 response.
Since Wolf’s arguments are old, we had
hesitated to respond. We do so now
because his analyses and his critique of our analyses are incorrect, and
because without such a refutation, his miscalculations, problematic evidence,
and dubious interpretation serve other scholars in a larger debate who continue
to define the pre-transition Chinese demographic regime in terms of the
Malthusian positive check (Brenner and Isett 2002; Cao and Chen 2002; Huang
2002).[8] The stakes, as proclaimed by Wolf, are
high since they have implications not only for China, but also for our
understanding of larger comparative social economic processes.
In this reply we focus on the three central issues
of Wolf's refashioned critique: the levels of pre-transition marital fertility,
the patterns of reproductive behavior behind them, and the use of poverty to
account for low Chinese fertility within marriage. We clarify his misrepresentations of our data, findings, and
interpretations. At the same time,
we illustrate how he relies on sources that are sometimes selective and other
times irrelevant, and uses demographic methods that are occasionally
inappropriate. Finally, we
demonstrate how in accounting for low marital fertility in China, Wolf relies
on an outdated understanding of the association between nutritional status and
fertility.
Pre-Transition Chinese
Marital Fertility
Wolf’s second objection, that
our data are flawed, ignores the measures we take to define and deal with their
limitations. [11] For instance, he notes our estimate
that the Liaoning household archives from Daoyi omit one-third of male births
and two-thirds female births (Lee and Campbell 1997, 13, 66), but does not
acknowledge our caveats about these data or our adjustments for this
undercount. After presenting our final estimate of a total marital
fertility rate (TMFR) of 6.3 for three thousand married Han bannermen from
Daoyi in Liaoning, he goes on to assert without explanation, "it is likely
that they considerably underestimate Chinese fertility. The true rate was at least 7.4 live
births per woman, which was the figure for the 71 women included in
Dickinson’s study of Jianyang." (136).[12]
Conversely,
Wolf does not address potential problems with his own data. While he cites results from his and
other analyses of mid-twentieth-century Taiwanese household registers, his own
survey of 580 women in 1980-81, several genealogy-based studies, and
Dickinson’s and several other small-scale Republican-era studies, he
neither discusses the selection of these samples nor the limitations of the
sources.[13] He seems unaware of any possible
survivor or cohort bias in his survey of elderly women,[14]
and does not acknowledge that the genealogy-based studies he cites often
exclude childless women and women who did not bear sons,[15]
and are accordingly biased upward.[16]
Wolf, moreover,
excludes several of his own findings that contradict his claims.[17] Thus while he insists that the Chinese
total marital fertility rate was between 7 and 8 (2001, 137), he does not
mention that his own estimates of total marital fertility in Haishan, Taiwan in
1906-10, 1911-15, 1916-20, 1921-25, 1926-30 were 6.25, 6.38, 6.15, 6.69, and
7.08 respectively before peaking at 7.41 and 7.94 in 1931-35 and 1936-40 (1984,
455). Putting aside the bottom
line, that even these numbers are inflated, and once recalculated hardly differ
from Barclay et al. (1976) and our own, Wolf’s use of early
twentieth-century fertility estimates from colonial Japanese Taiwan to
challenge fertility estimates from eighteenth and nineteenth century northeast
Qing China seems irrelevant.
Finally, Wolf’s claim of high pre-transition fertility depends not only on his selective use of data, but also his idiosyncratic calculation of total marital fertility. While demographers when comparing marital fertility typically either compare age-specific rates,[18] or a summary measure of total marital fertility for women age 20-49, excluding those age 15-19,[19] Wolf includes this age group and accordingly skews his comparisons.[20] To demonstrate the degree of this bias, we reexamine the age-specific rates on which he bases his claims. In Table 1, we compare two different summary measures of fertility, the total fertility rate (TFR) and the total marital fertility rate (TMFR) including and excluding women age 15 to 19. [21] Excluding females age 15-19 would lower the TFR for Haishan, Taiwan between 1906 and 1945 by 9 to 13 percent and for the seven locales of Wolf’s 1980-81 survey of elderly women by 6 to 12 percent. Excluding married females age 15 to 19 reduces the TMFR even more, by over 20 percent for Wolf's Haishan data, and by 18 percent for his 1980-1981 survey data, lowering the TMFR from Wolf’s 7.5 to somewhere between 4.9 and 6.2, which is consistent with our and other estimates of Chinese pre-transition marital fertility.[22]
Table 1 here
Pre-transition Chinese
Reproductive Behavior
The real
debate, of course, is not about the level of fertility but about its
determinants. Our own view is that
the low level of marital fertility in historical China at least partially
reflected deliberate control. We
say partially because low marital fertility in China resulted from a number of
behaviors, some of which reflected deliberate efforts to delay or forego
births, some of which had other purposes.
While couples were capable of
deliberately adjusting their fertility to their circumstances, they also
engaged for other reasons in other behaviors that lowered fertility such as low
coital frequency and very prolonged breastfeeding. Thus while we have emphasized deliberate control because its
presence is the most distinctive feature of Chinese marital fertility (Lee and
Wang 1999a and 1999b; Wang, Lee and Campbell 1995) we also recognized the
contribution of these other behaviors to China’s low pre-transition
marital fertility.
This
differs radically from Wolf’s (2001) characterization of our account of
low marital fertility. [23] We do not, as
Wolf suggests, claim that the relatively low marital fertility of the Chinese
was solely due to “birth
control,” that is, deliberate behavior specifically aimed at delaying or
foregoing births (2001, 137).[24] While we argue that for Chinese
couples, unlike European couples, control over reproduction was within the
‘calculus of conscious choice’ (Coale 1973), we do not claim that
all couples exercised control at all times. Nor do we claim that the differences in pre-transition
Chinese and European marital fertility were the product of parity-specific
birth control.
We
do not, as Wolf suggests, ‘accept as a
fact about Chinese behavior an elite medical recommendation that “coital
frequency should be no greater than three times a month for young adults, less
than twice a month for middle aged adults, and once a month at most for the
elderly” (Wolf 2001, 144). We
did not claim that all or even most Chinese obeyed such advice, at least not to
the letter. [25] Rather, we sought to show that such
advice was part of the context in which individuals made decisions. We accordingly noted that similar
recommendations were “by the eighteenth century [a] long established
consensus in the medical literature,” and that as with modern medical
advice against smoking, drinking, obesity, and lack of exercise “many
Chinese took this advice to heart” (Wang, Lee, Campbell 1995, 398).
In
asserting that our evidence of low coital frequency consists solely of
qualitative evidence from medical texts, Wolf (2001, 144) ignores the
quantitative evidence we cited in Wang, Lee, and Campbell (1995) and
misrepresents our explanation for low coital frequency.[26] Surveys of at least one contemporary
Asian population, Thailand, reveal that coital frequency for married couples is
substantially lower than in the United States.[27] Such direct dramatic differences in
sexual behavior between contemporary populations is hard to reconcile with
Wolf’s claim that lower coital frequency would require belief that the
Chinese were a “peculiarly asexual people” (Wolf 2001, 144).
Wolf’s discussion of the thirteen hypotheses that he claims to identify in our work is also problematic (2001, 145-151). For example, he treats purposive fertility behavior as evidence against the existence of control.[28] In particular, he interprets the low fertility of parents who have had sons relative to parents who have had daughters as evidence against fertility control (Wolf 2001, 146-147). He reasons that the higher fertility of parents with daughters reflects their anxiety over their failure to produce a son, and is indicative of a deliberate effort to increase fertility. We regard such purposive behavior as but one more form of control, since it is completely consistent with our view that Chinese couples sought to regulate their fertility, increasing it in some situations and decreasing it in others. Efforts to accelerate childbearing would distinguish Chinese couples from their European counterparts, for whom reproduction, if not ‘up to God’, was outside the ‘calculus of conscious choice.’ (Coale 1973).
Coale (1984), in his response to Wolf almost twenty years ago, pointed out that Wolf’s invocation of malnutrition to account for low marital fertility was not supported by empirical evidence from developing countries. [30] Summarizing results from studies that had recently been carried out in Guatemala and Bangladesh, Coale concluded, “if there is actual starvation, menstruation ceases and fertility is greatly reduced, but long-lasting malnutrition above the level of starvation seems to have very little effect on fertility” (477). In light of this, Wolf’s (2001) continued invocation of malnutrition to account for low marital fertility appears obstinate. He has not cited any scientific evidence that contradicts Coale (1984, 477). Since Coale’s discussion of the relationship between nutrition and fertility is almost two decades old, we conclude our reply with a review of recent, relevant findings.
While poor nutrition delays the onset of menstruation, even in poorly nourished populations the age at menarche was low enough that it was marginal to comparisons of marital fertility between Europe and China, which should exclude women age 15 to 19. Even the late ages of menarche that Wolf (2001, 140) claims for China in the past, 16 to 17, were lower than the average age of female marriage in the historical Chinese populations for which data are available summarized in Lee and Wang (1999b, 67). In Liaoning, it was 18 Western years of age (Lee and Campbell 1997, 86). In the Qing imperial lineage, it was between 20 and 21 (Lee and Wang 1996). For Chinese women born between 1900 and 1925, according to results from the nationally representative Two-per-Thousand survey it was 19.0 (Wang and Tuma 1993). In Taiwan during this same period, it was if anything slightly higher (Barclay 1954, 211).[31] We accordingly doubt that delayed menarche of the sort that Wolf suggests would have had a substantial effect on the fertility of married couples.
Even for women who had already reached menarche, the effects of poor nutrition on their fertility are likely to have been mild at most. A series of studies carried out in developing countries since the late 1970s have concluded that chronic poor nutrition by itself does not induce amenorrhea, and therefore does not account for low fertility (Bongaarts 1980; John, Menken, and Chowdhury 1987; Menken, Trussell, and Watkins 1981). Only severe, acute malnutrition of the type observed during famines induces amenorrhea. Most recently, a study in the Philippines found that once women who had been pregnant resumed menstruating, nutritional status did not increase subsequent time to conception (Popkin et al. 1993).
As for intra-uterine mortality, malnutrition does raise its likelihood substantially, but the net effect on fertility should have been mild. The baseline chances of fetal death are low enough that even doubling them has only a small proportional effect on the chances of carrying a pregnancy to term. According to a study carried out in Matlab, Bangladesh, the women who weighed the least had twice the chances of intra-uterine mortality of women whose weight was comparable to contemporary American women of the same height: 6 percent versus 3 percent (Pebley, Huffman, Chowdhury, and Stupp 1985, 438). Even with a generous allowance for a much higher baseline risk, say 10 percent of observed pregnancies (Leridon 1977, 63), the implication is that moving from one tail of the distribution of nutritional statuses to the other would only reduce the chances of carrying a pregnancy to term from 90 percent to 80 percent. Moreover, a fetal death averts less than one birth, because women are eligible to conceive again soon afterward.
While there is scientific evidence that postpartum amenorrhea is longer for poorly nourished women than for well-nourished ones, it is unlikely that such an effect could have accounted for the gap between European and Chinese birth intervals that even Wolf (2001, 142) acknowledges. According to a study of women in the Philippines, postpartum amenorrhea was only four months longer for poorly nourished women than for well-nourished women (Popkin et al. 1993). A study in Matlab, Bangladesh yielded similar estimates of the difference between well- and poorly-nourished women in the length of postpartum amenorrhea (Huffman et al. 1987). Neither of these effects would account for the differences between European and Chinese birth intervals of a year or more.
As for Wolf’s claim that poor nutrition may have induced early menopause, there is no conclusive evidence that such an effect could have accounted for the low average age at last birth. While results from some studies suggest that nutrition may affect menopause, for example, some studies have shown than thinner women reach menopause earlier than heavier women, the strength and source of the relationship remains unclear (Rahman and Menken 1993, 67). Even if nutrition had an effect on the age at menopause, however, it probably would not have accounted for the relatively low at last birth in historical China. Studies in both developed and developing countries almost all report an average age at menopause of around 50 years. The lowest average age reported in the studies surveyed by Rahman and Menken (1993, 66) was 43.6, for a population of malnourished Melanesians living in New Guinea. This is still several years higher than the mean at age at last birth that Wolf accepts for China, 38 or 39.
We are skeptical that the other, non-physiological mechanisms that Wolf invokes to account for low Chinese fertility were more important in China than in Europe. For example, Wolf suggests spousal separation as a result of migration to explain low marital fertility in China. We do not believe there is any a priori reason to expect that spousal separation was more common in historical China than in historical Europe, or that peasants worked harder. Indeed, in two of the Chinese populations where we have observed low marital fertility, spousal separation is highly unlikely to have played a role: the members of the Qing imperial lineage were confined to Beijing (Lee, Campbell, and Wang 1993), and the residents of the Liaoning state farms could only migrate legally within the state farm system (Lee and Campbell 1997).
Even
if Wolf’s (2001) iconoclastic views about the influence of nutrition on
fertility were correct, it is by no means clear that the Chinese were as poor
and malnourished as he claims.
While Wolf (2001) takes as a given the grinding poverty of the Chinese
before 1949, there is in fact a vigorous debate over the comparative standards
of living in Europe and China, especially before the nineteenth century (Fang
1996; Lee and Wang 1999b; Li 1998; Pomeranz 2000). Although this debate is by no means settled (Brenner and
Isett 2002; Huang 2002; Pomeranz 2002), in light of the evidence presented by
us and others it does seem premature to invoke the extreme poverty of the
Chinese to explain any distinct features of their demographic behavior. This, of course, is especially true
when the Chinese under discussion are the Imperial nobility.
|
Table 1 Fertility rates with and without age group 15-19 |
|||||||||||
|
|
|
|
|
|
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
|
|
||
|
Wolf's Hai-shan, Taiwan
data, 1906-45 |
|
|
|
|
|
|
|||||
|
|
|
|
|
|
|
|
|
|
|
||
|
|
Age-Specific Fertility |
|
|
|
|||||||
|
|
|
|
|
|
|
|
|
|
|
||
|
Period |
15-19 |
20-24 |
25-29 |
30-34 |
35-39 |
40-44 |
Total |
w/o 15-19 |
% Difference |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
1906-10 |
139 |
252 |
231 |
212 |
157 |
86 |
5.4 |
4.7 |
13 |
||
|
1911-15 |
123 |
250 |
229 |
206 |
173 |
84 |
5.3 |
4.7 |
12 |
||
|
1916-20 |
111 |
252 |
230 |
206 |
159 |
74 |
5.2 |
4.6 |
11 |
||
|
1921-25 |
128 |
267 |
251 |
207 |
159 |
76 |
5.4 |
4.8 |
12 |
||
|
1926-30 |
139 |
271 |
261 |
222 |
170 |
78 |
5.7 |
5.0 |
12 |
||
|
1931-35 |
130 |
255 |
267 |
240 |
175 |
82 |
5.7 |
5.1 |
11 |
||
|
1936-40 |
114 |
261 |
270 |
251 |
196 |
93 |
5.9 |
5.4 |
10 |
||
|
1941-45 |
88 |
222 |
236 |
208 |
163 |
89 |
5.0 |
4.6 |
9 |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
|
Age-Specific Marital Fertility |
|
|
|
|||||||
|
|
|
|
|
|
|
|
|
|
|
||
|
Period |
15-19 |
20-24 |
25-29 |
30-34 |
35-39 |
40-44 |
Total |
w/o 15-19 |
% Difference |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
1906-10 |
261 |
263 |
237 |
221 |
171 |
98 |
6.3 |
5.0 |
21 |
||
|
1911-15 |
271 |
271 |
243 |
209 |
186 |
95 |
6.4 |
5.0 |
21 |
||
|
1916-20 |
245 |
281 |
239 |
212 |
169 |
84 |
6.2 |
4.9 |
20 |
||
|
1921-25 |
296 |
306 |
266 |
222 |
160 |
78 |
6.6 |
5.2 |
22 |
||
|
1926-30 |
310 |
318 |
282 |
239 |
184 |
84 |
7.1 |
5.5 |
22 |
||
|
1931-35 |
336 |
311 |
291 |
259 |
188 |
96 |
7.4 |
5.7 |
23 |
||
|
1936-40 |
353 |
337 |
305 |
275 |
213 |
105 |
7.9 |
6.2 |
22 |
||
|
1941-45 |
321 |
316 |
283 |
243 |
182 |
100 |
7.2 |
5.6 |
22 |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
Wolf's 1980-81 survey |
|
|
|
|
|
|
|
|
|||
|
|
Age-Specific Fertility |
|
|
||||||||
|
|
|
|
|
|
|
|
|
|
|
||
|
Location |
15-19 |
20-24 |
25-29 |
30-34 |
35-39 |
40-44 |
Total |
w/o 15-19 |
% Difference |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
Beijing |
142 |
312 |
312 |
254 |
165 |
27 |
6.1 |
5.4 |
12 |
||
|
Fujian |
128 |
238 |
293 |
253 |
158 |
55 |
5.6 |
5.0 |
11 |
||
|
Zhejiang |
119 |
294 |
326 |
205 |
173 |
49 |
5.8 |
5.2 |
10 |
||
|
Jiangsu |
77 |
313 |
283 |
265 |
160 |
54 |
5.8 |
5.4 |
7 |
||
|
Shandong |
67 |
264 |
269 |
240 |
196 |
80 |
5.6 |
5.2 |
6 |
||
|
Shanxi |
163 |
261 |
268 |
226 |
161 |
54 |
5.7 |
4.9 |
14 |
||
|
Sichuan |
69 |
251 |
302 |
273 |
167 |
118 |
5.9 |
5.6 |
6 |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
All |
106 |
275 |
298 |
246 |
169 |
65 |
5.8 |
5.3 |
9 |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
Age-Specific Marital Fertility |
|||||||||||
|
|
|
|
|
|
|
|
|
|
|
||
|
Location |
15-19 |
20-24 |
25-29 |
30-34 |
35-39 |
40-44 |
Total |
w/o 15-19 |
% Difference |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
Beijing |
261 |
348 |
335 |
279 |
183 |
30 |
7.2 |
5.9 |
18 |
||
|
Fujian |
346 |
279 |
344 |
294 |
205 |
78 |
7.7 |
6.0 |
22 |
||
|
Zhejiang |
244 |
331 |
349 |
224 |
190 |
55 |
7.0 |
5.7 |
18 |
||
|
Jiangsu |
285 |
347 |
294 |
277 |
175 |
60 |
7.2 |
5.8 |
20 |
||
|
Shandong |
249 |
334 |
282 |
255 |
216 |
91 |
7.1 |
5.9 |
17 |
||
|
Shanxi |
241 |
265 |
269 |
230 |
168 |
58 |
6.2 |
5.0 |
20 |
||
|
Sichuan |
185 |
274 |
313 |
293 |
190 |
148 |
7.0 |
6.1 |
13 |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
All |
255 |
309 |
308 |
264 |
189 |
76 |
7.0 |
5.7 |
18 |
||
|
|
|
|
|
|
|
|
|
|
|
||
|
|
|
|
|
|
|
|
|
|
|
||
|
Sources: Hai-shan, Taiwan,
Wolf 1985, 454, 455; Wolf 1980-81 survey, Wolf 1985, 458, 459. |
|||||||||||
REFERENCES
Barclay, George W. 1954. Colonial Development and Population in Taiwan. Princeton: Princeton University Press.
Barclay, George W., Ansley J. Coale, Michael A. Stoto, and
James Trussell. 1976. ‘A reassessment of the demography
of traditional rural China.’
Population Index 42: 606-35.
Bongaarts, John.
1980. “Does
malnutrition affect fecundity? A
summary of the evidence.” Science. 208:564-69.
Brenner, Robert and Christopher Isett. 2002. “England
divergence from China’s Yangzi Delta: property relations, microeconomics
and patterns of development.” Journal of Asian Studies 61(2):.
Cao Shuji and Chen Yixin. 2002.
“Maersasi lilun he Qingdai yilaide Zhongguo renkou: ping Meiguo
xuezhe jinnianlai de xiangguan yanjiu” (Malthusian theory and Chinese population since the Qing: A
critique of recent American scholarship).
Lishi Yanjiu (Historical
research) 1: 41-54.
Coale, Ansley J.
1973. “The
Demographic transition reconsidered.” In Proceedings of the International Population
Conference. Vol. 1. Liège: International Union for
the Scientific Study of Population, 58-71.
_____1980.
“Fertility in prerevolutionary rural China: In defense of a
reassessment.” Population
and Development Review. 10: 471-480.
_____1985. “Fertility in rural China: A reconfirmation of the Barclay reassessment.” In Hanley and Wolf, eds. 86-195.
Coale, Ansley J. and Chen Shengli. 1987. Basic Data on Fertility in the Provinces of China,
1942-1982. Honolulu: East-West Center.
Coale, Ansley J. and James Trussell. 1974. “Modeling fertility schedules: Variations in the
structure of childbearing in human populations.” Population Index. 40: 203-213.
Fang,
Xing 1996. ‘Qingdai jiangnan nongmin de xiaofei’ (On the
consumption of peasants in Qing Jiangnan), in Zhongguo jingjishi yanjiu (Researchers on Chinese Economic History) (Beijing),
3: 91-98.
Frisch, Rose.
1978 “Population,
food intake, and fertility.” Science
199: 22-30.
Gray, Ronald H.
1983. “The impact of
health and nutrition on natural fertility.” In Rodolfo Bulatao and Ronald Lee, eds. Determinants of Fertility in
Developing Countries. Vol. 2, 139-62.
Hajnal, John. 1965. “European marriage patterns in perspective.” In Glass and Eversley, 101-146.
_____1982. “Two kinds of pre-industrial household formation system.” Population and Development Review 8: 449‑494.
Hanley, Susan and Arthur P.
Wolf. Eds. 1985. Family and Population in East Asian
History. Stanford: Stanford University Press.
Harrell, Stevan.
Ed. 1995. Chinese Historical Microdemography. Berkeley:
University of California Press.
Huang, Philip.
2002. “Development or
involution in Eighteenth-Century Britain and China? A Review of Kenneth Pomeranz’s The Great Divergence:
China, Europe, and the Making of the Modern World Economy.” Journal of Asian Studies. 61: 501-538.
Huffman, Sandra L., Kathleen Ford, Hubert A. Allen, and
Peter Streble. 1987. “Nutrition and fertility in
Bangladesh: Breastfeeding and postpartum amenorrhea.” Population Studies. 41: 447-462.
John, A. Meredith, Jane A. Menken, and A. K. M. Alauddin
Chowdhury. 1987. “The effects of breastfeeding and
nutrition on fecundability in rural Bangladesh: A hazards-model
analysis.” Population
Studies. 41: 433-46.
Laslett, Peter. 1983. “Family and household as workgroup and kin group: areas of traditional Europe compared.” In Wall, Robin, and Laslett, 513-563.
_____1988. “Family, kinship and collectivity as systems of support in pre-industrial Europe: a consideration of the ‘nuclear hardship’ hypotheses.” Continuity and Change 3: 153-175.
Lavely, William R. 1986. “Age patterns of Chinese marital fertility,
1950-1981.” Demography. 23: 419-434.
Lee, James and Cameron Campbell. 1997. Fate and Fortune in Rural China: Social
Organization and Population Behavior in Liaoning, 1774-1873. Cambridge:
Cambridge University Press.
Lee, James, Cameron Campbell, and Wang Feng. 2002. “Positive Checks or Chinese Checks?” Journal of Asian Studies. 61: 591-607.
Lee, James and Wang Feng. 1999a. “Malthusian models
and Chinese realities: The Chinese demographic system 1700 – 2000.”
Population and Development Review 25:
33-65.
Lee, James and Wang Feng. 1999b. One Quarter of Humanity: Malthusian Mythology and Chinese
Realities 1700-2000. Cambridge: Harvard
University Press.
Lee, James, Wang Feng, and Cameron Campbell. 1994.
“Infant and child mortality among the late imperial Chinese nobility:
implications for two kinds of positive check.” Population Studies 48: 395-411.
Leridon, Henri.
1977. Human Fertility:
The Basic Components. Chicago: University of Chicago Press.
Li, Bozhong.
1998. Agricultural
Development in Jiangnan. 1620-1850. New York:
St. Martin’s Press.
Liu, Ts’ui-jung. 1992. Ming
Qing shiqi jiazu renkou yu shehui jingji bianqian (Lineage population and socio-economic changes in the Ming
and Qing periods). Taibei:
Institute of Economics, Academia Sinica. 2 vols.
_____1995. ‘Historical demography of South China
lineages.’ In Harrell 1995,
94-120.
Macfarlane, Alan. 1978. The Origins of English Individualism: Family, Property, and Social Transition. Oxford: Oxford University Press.
_____1986. Marriage and Love in England, Modes of Reproduction 1300-1840. Oxford: Basil Blackwell.
_____1987. The Culture of Capitalism. Oxford: Oxford University Press, 1987.
_____1997. The Savage Wars of Peace: England Japan and the Malthusian Trap. Oxford: Basil Blackwell.
Menken, Jane A., James Trussell, and Susan Watkins. 1981. “The nutrition-fertility link: An evaluation of the
evidence.” Journal of
Interdisciplinary History. 11: 425-41.
Pebley, Anne R., Sandra L. Huffman, A.K.M. Alauddin
Chowdhury, and P.W. Stupp.
1985. “Intra-uterine
mortality and maternal nutritional status in rural Bangladesh.” Population Studies. 39: 425-440.
Pomeranz, Kenneth.
2000. The Great
Divergence: China, Europe, and the Making of the Modern World Economy. Princeton:
Princeton University Press.
_____2001.
“Beyond the East-West binary: resituating development paths in the
eighteenth-century world.” Journal
of Asian Studies 61: 539-590.
Popkin, Barry M., David K. Guilkey, John S. Akin, Linda S.
Adair, J. Richard Udry, and Wilhelm Flieger. 1993.
“Nutrition, lactation, and birth spacing in Filipino women.” Demography. 30: 333-52.
Rahman, Omar and Jane Menken. 1993.
“Age at menopause and fecundity preceding menopause.” In Gray, Ronald with Henri Leridon and
Alfred Spira, eds. Biomedical
and demographic determinants of reproduction. Oxford, England: Clarendon Press,
65-84.
Rindfuss, Ronald and Philip Morgan. 1983. “Marriage, sex, and the first birth interval: The
quiet revolution in Asia.” Population
and Development Review. 9: 259-278.
Telford, Ted A. 1995. ‘Fertility and population growth in the lineages of Tongcheng County 1520-1661.’ In Harrell 1995, 48-93.
Wang, Feng and Nancy B. Tuma. 1993. “Changes in Chinese marriage patterns during the twentieth century.” Proceedings of the IUSSP International Population Conference. Montreal. 337-352.
Wang, Feng and Quanhe Yang. 1996. “Age at marriage
and the first birth interval: The emerging change in sexual behavior among
young couples in China.” Population
and Development Review. 22: 299-320.
Wang Feng, James Lee, and Cameron Campbell. 1995. “Marital fertility control among the Qing nobility: implications for two types of preventive check.” Population Studies 49.3 (November): 383-400.
Wang Feng and James Lee. 2002. “Zhaidiao renkou juedinglun de guanghuan” (Correcting population determinism). Lishi yanjiu (Historical research) 1: 57-63.
Wolf, Arthur P.
1984. “Fertility in
prerevolutionary rural China.”
Population and Development Review. 10: 443-470.
_____1985. “Fertility in pre-revolutionary
rural China.” In Hanley and
Wolf, 154-185.
_____2001.
“Is there evidence of birth control in late imperial
China?” Population and
Development Review. 27: 133-154.
Wolf, Arthur and Chieh‑shan Huang. 1980. Marriage and Adoption in China, 1845‑1945. Stanford: Stanford University Press.
Wrigley, Edward A., and Roger Schofield. 1981. The Population History of England, 1541-1871: a reconstruction. Cambridge: Cambridge University Press.
Wrigley, Edward Anthony, Ros Davies, James Oeppen, and
Roger Schofield. 1997. English Population History from
Family Reconstitution, 1580-1837. Cambridge: Cambridge University Press.
Yao Xinwu and Yin Hua. 1994. Zhongguo
changyong renkou shuju ji (A collection of
basic data on China’s population).
Beijing: Zhongguo renkou chubanshe.
Zhao Zhongwei.
1997. “Deliberate
birth control under a high-fertility regime: Reproductive control in China
before 1970.” Population
and Development Review. 23: 729-67.
Zheng Zhenman. 2000. “Jinbainian Mindong yanhai di hunyin, jiating, yu shengyulü – Lianjiang xian Pukouzhen Guanlingcun diaocha baogao (Fertility, marriage and family on the Fujian coast during the last century – a research report on Lianjiang County, Pukou zhen, Guanling Village),” in Li Zhongqing, Guo Songyi, and Ding Yizhuang. Eds. Hunyin, jiating, yu renkou xingwei: Dongxi bijiao (Marriage, family formation, and population behavior: East-West comparisons). Peking: Peking University Press, 61-81.
[1]
Wolf initially presented his criticism of Barclay et al. (1976) in 1978
and published his exchange with Coale in Population and Development Review in 1984 as well as in Hanley and Wolf
(1985).
[2] Notably Stevan Harrell, William Lavely,
Ts-ui-jung Liu, Ted Telford, Arthur Wolf, and Zhao Zhongwei among others. See Lee and Wang 1999b for a complete
bibliography.
[3] In Lee and Wang (1999) we surveyed
studies of some 500 thousand individuals who lived before 1950. Continuing
research during the last four years has already added another 300 thousand
individuals and should continue to add many more, broadening our understanding
of Chinese historical population processes.
[4]
See Lavely and Wong (1998) and Zhao (1997) for two other comparisons using
Chinese data, and Das Gupta 1995 for a comparison of the European and Indian
experience.
[5]
By “system” we mean the defining characteristics of Chinese
demographic behavior during the last 300 years in contrast to the “European
demographic system” identified by Flinn (1981).
[6]
“The revisionists and I agree that in China marriage was early and
nearly universal for females; we agree that as compared with premodern Europe,
marital fertility was moderate; that birth intervals were longer in China than
in Europe; that in China the interval from marriage to first birth was
particularly long; that marital fertility in China followed a natural fertility
trajectory; that it was higher among the wealthy than the poor; that the Chinese
used sex-selective infanticide to regulate family size; that they used both
make and female adoption for the same purpose; that most Chinese recognized the
relationship between breastfeeding and child spacing; that they used this
knowledge to achieve fertility goals; and most importantly, we agree that their
attitude toward reproduction was eminently rational (Wolf 2001, 134).”
[7] Wolf argues
that any low marital fertility in China was an involuntary consequence of
“poverty,” and that low marital fertility is therefore
“evidence of chronic malnutrition, untreated diseases, hard manual labor,
and economically-enforced conjugal separation” (Wolf 2001, 137).
[8] We address elsewhere (Lee, Campbell, and
Wang 2002; Wang and Lee 2002) the problems with these Malthusian
interpretations of Chinese history.
[9] Wolf’s (2001) apparent belief that
the level of fertility can be used to reveal the presence or absence of
fertility control is iconoclastic.
It is precisely because the level of fertility is unreliable as an indicator
of the presence of fertility control that Coale and Trussell (1974) developed
procedures for detecting parity-specific control from the age pattern of
fertility and that we identify other patterns of fertility behavior in
Wolf’s list of our thirteen hypotheses (2001, 145).
[10] They include Liu (1995) and Telford
(1992) for a variety of other Chinese lineage populations during the Ming and
Qing, estimates by Barclay et al. (1976) for 22 provinces of China in 1929-31,
Wolf’s (1985) own estimates of age-specific marital fertility rates in
Haishan, Taiwan in 1906-1910, and estimates for China as a whole after 1949 by
Coale and Chen (1987), Lavely (1986, 432-33), and Yao and Yin (1994).
[11] Wolf (2001) dismisses the genealogical
archives of imperial lineage because even though these data can be
“accepted as accurate” the population was unrepresentative of
China. Conversely, he dismisses
the household register archives of royal peasants in Liaoning because even
though “they were representative of the larger population,” the
data are inaccurate (136).
[12] This is but one example of a preference
for assertion and illustration over deduction or induction as evidence. Wolf, for example, speculates that the
registers were inaccurate because “it seems likely that some young men
avoided registration by bribery or flight” (136), ignoring our analyses
of data quality and under-registration from mortality
patterns in the data and comparison with model life tables (Lee and
Campbell 1997, 65-70, 223-237). He applies a similar logic in
rejecting what he claims are our estimates for the lengths of the first birth
interval, the intervals between births, and the age at last birth.
[13] Many of these samples are not only
small, but potentially unrepresentative.
While Wolf’s own interviews of elderly women may suggest that they
sought to have as many sons as possible, other retrospective surveys of
different elderly women have come to opposite conclusions. Zheng’s (2000) survey of Guanling
village in Fujian Province is a good example. According to this 1994-1995 study of 50 women aged
over 60, 37 reported that they did not want more children (70-71).
[14]
Wolf’s 1980-81 survey only included elderly women who had spent
the majority of their
reproductive years in the pro-natalist period of 1949-1969. Fertility estimates based on these
women’s experiences may be biased upward because those fortunate enough
to survive to old age are likely to have been healthier and had more children
than those who died at younger ages.
[15] Referring to the estimates of marital
fertility from lineage genealogies, Liu (1995, 100) notes that they were
“derived from age-specific fertility rates, which required the data from
each conjugal family to include quite complete vital dates for each members,
and those families with no sons were not included in the observation.”
[16]
Wolf compounds such biases when he
multiplies the rates reported by Liu (1995) by 1.5 to account for
under-registration (2001, 137).
While an adjustment may be in order, the one he
applies appears arbitrary, as he provides no explanation for choosing it over
other possible values.
[17]
Similarly, when
Wolf (2001, 137) lists 11 estimates of marital fertility for rural communities
in Taiwan, two of which, 7.61 and 7.78, he says are from Wolf (1995, 120), he
fails to mention that this source presents six estimates of the level of
marital fertility for different population subgroups depending on ethnicity
(Hakka or Hokkien) and type of marriage (major, minor, or uxorilocal). The 7.61 in Wolf (2001, 137) seems to
correspond to Hokkien in major marriages.
We could not match 7.78 to any of the estimates in Wolf (1995, 120), the
highest of which was 7.69, for Hakka in major marriages. The total marital fertility rates for
Hokkien and Hakka in minor marriages are much lower, 6.02 and 6.19,
respectively (Wolf 1995, 120).
[18]
Coale (1984, 475) in his response to Wolf (1984) plots age-specific marital
fertility rates from Haishan Taiwan 1905-1945 and from Barclay et al. (1976) to
show that they are almost identical to each other and appreciably lower than
historic European marital fertility rates. We do the same in our comparison of age-specific marital
fertility rates from a variety of pre-transition East Asian and West European
populations (Lee and Wang 1999b, 87),
[19] Thus the calculation of TMFR by Flinn for pre-transition European populations (1981, 31), and by Wrigley, Davies, Oeppen, and Schofield (1997, 427) for selected pre-transition English populations, only include married women age 20-49 as do the comparisons by Lee and Wang (1999, 87) for historical Chinese populations, and by Lavely (1986) for contemporary Chinese populations.
[20]
Including married women age 15-19 would raise the TMFR for
pre-transition English populations to almost 12 (Wrigley, Davies, Oeppen, and
Schofield (1997, 416)
[21] The total marital fertility rate (TMFR) that Wolf
relies on is a summation of age-specific marital fertility rates. It reflects the number of births a
woman would have if she married by the beginning of the earliest age group and
remained married until the end of the last age group. Thus a TMFR based on age-specific rates from 15 to 49 would
reflect the number of children a woman would be expected to have if she married
at age 15 and remained married until age 49. A TMFR based on rates from ages 20 to 49 would reflect the
number of births expected for a woman who married at age 20. Comparisons of TMFR can be
misleading if they include ages at which few women are married, in particular,
the age group 15 to 19. Relatively
few women are married in this age group, and the marital fertility of those who
are married tends to be unusually high.
Including this age group generates a misleading impression that every married woman would have the same number of births as
those who married as young as 15 to 19.
Most calculations and comparisons of the TMFR accordingly begin at age
20.
[22]
The total marital fertility rates for genealogy populations from Anhui
presented in Telford (1995, 50) are similarly inflated as he too includes
married women age 15 to 19 in his calculations.
[24] “The reason [for low marital
fertility in China], they say, was birth control” (Wolf 2001, 137).
[25] In Lee and
Wang (1999b, 91) we concluded that “the low fertility and long birth
intervals of Chinese couples in the past were at least in part the result of
their ability and even willingness to regulate coital frequency.”
[26] This is true for our discussions of
early starting as well as long birth intervals. Whereas according to Wolf (2001, 140) we claim that
intervals between marriage and first birth were long because married couples
“deliberately delayed having children,” we actually attributed long
intervals at least partly to low coital frequency. In particular, in our discussion of this phenomenon we
referred to more recent studies of long first birth intervals in Asia that
suggested a role for the lack of conjugal passion that may have characterized
arranged marriages, at least in their early stages (Rindfuss and Morgan 1983;
Wang and Yang 1996). Thus we noted
that late starting was common among the two populations we have analyzed in
detail and conclude in Lee and Campbell (1997, 93) that “long intervals
between marriage and first birth were until recently common throughout China
and much of Asia and are usually attributed to the relatively low coital
frequencies assumed to have characterized arranged marriages.”
[27] ‘Fertility
surveys reveal that even today when couples have the protection of
contraception, Asian couples continue to follow a pattern of coital frequency
considerably lower than elsewhere.
In Thailand, for example, the mean coital frequency of all currently
married women during the four weeks preceding the 1987 Demographic and Health
Survey was 3.2. Newlyweds only had
a monthly coital frequency of 6, which dropped to 4.2 after one year of
marriage and 3.7 after four years of marriage. See N. Chayovan and J. Knodel, “Coital activity among
married Thai women: evidence from the 1987 Thailand Demographic and Health
Survey.” Research Reports of
the Population Studies Center, University of Michigan, No. 91-221, 1991. The comparable number in the United
States in 1975 was 8.9 for all currently married women and 10.4 for women in
their first five years of marriage.
See J. Trussell and C. Westoff, “Contraceptive practices and
trends in coital frequency.”
Family Planning Perspectives.
Vol. 12, 1980: 246-249.”
(Lee, Wang, and Campbell 1995, 398).
[28]
Space and time constraints prevent us from
dealing with the remainder of Wolf’s (2001, 145-151) discussion of the
thirteen hypotheses he claims to identify in our publications. While
we encourage readers to read Wang, Lee, and Campbell (1995), Lee and Campbell
(1997) and Lee and Wang (1999a and 1999b), we would like to suggest that this
is an appropriate place for application of Occam’s Razor. When one simple and straightforward
explanation can plausibly account for thirteen distinct phenomena, parsimony
seems to dictate that it be favored over the unwieldy combination of separate
explanations, conflations, refutations, and dismissals that Wolf (2001,
145-151) offers.
[29]
“The revisionists are right in insisting that marital fertility
was lower in China than in parts of Western Europe and very much lower than in
such exemplary populations as the Hutterites. The reason, they say, was birth control. The reason, I say, was poverty.”
(Wolf 2001, 137)
[30]
This is not only the case with the level of pre-revolutionary fertility,
but also with Wolf’s repeated claim that any observed low Chinese
fertility was a consequence of malnutrition or poverty. As Coale (1984) replied, “There
is a theme running through much of the discussion in Wolf’s paper to the
effect that fertility is negatively associated with poverty and positively
associated with level of living.
As a general proposition, I think the assumption of a close association
between level of living and fertility in noncontraceptive population is
erroneous.” (477)
[31]
In Haishan, Taiwan between 1891 and 1921, median age at first marriage
ranged between 18.2 and 19.6 for women in ‘major’ marriages, 16.8
and 17.4 for women in ‘minor’ marriages, and 17.4 to 19.6 for women
in uxorilocal marriages (Wolf and Huang 1980, 135)