On a Small Death

Who is there to weep for your little girl who died too soon?

Yet we all wept with reasons each our own:

Grandma wept conventional tears

Through glad at heart burdened house was rid of a girl;

Mother - poor Mother - secretly shed her sorrow, thought you never knew;

And the good neighbors who always joint to help the dead to final rest,

Why should they weep for what is second-hand sorrow to them?

I could have wept, but thought: why weep for so small a death?

And thus, we all mourned, as men must at death,

According to custom, resting foreheads in our hands.

Finally we lifted you, a small burden, form that sad house;

We walked a little, turned the corner toward the burning grounds

And there she was, your little friend, at the window.

She looked intently, watching your new game:

To Climb up like that and sleep on grown-ups shoulders:

We, the mournful, paying no attention moved on, Suddenly it dawned upon your friend;

This cruel game - your grim departure: She cried her lament aloud.

She moaned, the only one in the world to feel your loss.

And I who never meant to weep

could hardly hold my tears.

- Uma Shankar Joshi

This poem, written by the Hindi poet Uma Shankar Joshi in the 1930s, focuses on the untimely deaths of daughters in northern India. The fact that small girls in northern India do not survive at the same rates as their brothers has been well known in academic circles for some years, as indeed it was by thoughtful Indians even earlier. It has been gaining wider publicity in the West due to the use of Western medical technology to achieve the same end - through amniocentesis, which then destroys them. Whether a small girl dies an untimely death through neglect and lack of medical care and food or dies as a fetus, both achieve the same end: there are fewer daughters for northern Indian families to feed, educate and marry off, a process usually requiring substantial dowries.

Since 1901, the all-India population has become increasingly male (with the female-to-male ratio dropping from 972 to 930). Moreover, child mortality has seen a major decline since the 1940s. These data imply that ex-specific mortality has continued, if not intensified, under conditions of better health care and socioeconomic change. An increasingly male population, a declining child mortality rate and a historical record of sex-specific juvenile mortality all point to an increasing disparity between female and male survival - that is, even though mortality is declining, more boys may be surviving than girls. This creates a problem in understanding fertility behavior; for here is a situation in which changes in that behavior are occurring in the context of this increase in sex-specific mortality.

Most problematic recent evidence from northern India that, amniocentesis aside, a young rural girl's chances of survival today in contrast to those of her brothers are even less than in the past. This growing bias against daughters is not specific: it appears in rich, peasant families and in poor, landless families. Although the immediate socioeconomic factors for this discrimination differ in these two groups, its roots derive from a variety of common structural factors in Indian culture that marginalize women.

A Study of Two Communities

Evidence of increasing selective discrimination against female children among rich, peasant farmers comes from a restudy by Monica Das Gupta of 11 villages in Ludhiana district surveyed in the 1950s in the Khanna Study. Ludhiana is located some 200 miles north of New Delhi, in the prosperous state of Punjab. Evidence of the increasing female bias in child mortality among the poor comes from my own work in the community of Karimpur, some 150 miles southeast of New Delhi in the rather backward district of Mainpuri, in the western part of the state of Uttar Pradesh.

Both these areas have been transformed by the Green revolution, although Ludhiana grew earlier and has achieved significantly greater prosperity than western Uttar Pradesh. In Ludhiana, the amenities of health services, schools, electricity and other aspects of the state infrastructure are well developed. In Mainpuri, schooling remains a privilege reserved for the elite, health services are inadequate, electricity has reached only a few homes and prosperity is beyond the reach of most. In Ludhiana, the relative affluence of the past two decades has reached landowners and landless alike, the former prospering from agricultural gains and the latter from new employment opportunities in industry. Mainpuri, however, has no industrial development. Although most farmers there have gained from the improved outputs wrought by the Green revolution, their situation remains precarious. Meanwhile, Mainpuri's landless have been forced out of agricultural employment and have only labor jobs in the district town as options.

These two areas have major demographic differences. Ludhiana is well into the third stage or the demographic transition: fertility and mortality rates have fallen steadily, with a crude birth rate in 1984 of 28 per thousand and an infant mortality rate of 62 per thousand for the period 1979-1984. Karimpur, however, has not yet entered the demographic transition: its fertility rate continues to rise (as is frequent in the early stage of modernization) and its mortality rate is still high. Between 1969 and 1974, the crude birth rate among the poor in Karimpur was 41.8 per thousand; it rose to 47.5 per thousand in the period 1975-1984. The comparable rates for the Brahman landlords were 35.1 per thousand and 40.1 per thousand in the period 1975-1984. The comparable rates for the Brahman landlords were 35.1 per thousand and 40.1 per thousand, respectively. Infant mortality rates have declined significantly in the past 50 years - from a village average of 334 per thousand in the period 1932-1951 to 202 per thousand in the period 1975-1984 - but they remain far above those in Ludhiana Child death is still a reality to Karimpur mothers: 82 percent of mothers who have had two or more children have seen at least one of their children die.

What is significant about both Ludhiana and Karimpur, however, is the growing disparity between male and female child deaths in certain groups. Tables 2 and 3 show sex differentials in infant and child mortality in these two communities. In the neonatal period, male mortality is greater than female mortality; at a time when biological factors are the dominant cause of death, this is to be expected. But after the first month of the life when factors such as quality of life, environment and care are more critical to life chances, female deaths outnumber male deaths. These factors, which can be controlled by human agency, reflect the pervasive societal value that lead to discrimination against female children.

Given the much greater mortality rates in Karimpur than in the Khanna Study - rates that, even in the 1980s, have not reached the Khanna levels of the 1950s - the male-to-female rations are not quite as marked in the later age groups as in Khanna, and both males and females are susceptible to death at all ages. When death is all-pervasive, as it essentially still is in Karimpur, it is somewhat harder to discriminate against female children (without outright female infanticide) - although census data from the 1880s onward show that females certainly had fewer life chances than did males all over northern India. In Karimpur, everyone's life chances are well below those living in Ludhiana. Even with the greater overall mortality, however, girls older than one month have less chance of survival than do their brothers. The ratio of male of female mortality for the most recent period for Karimpur, 1972-1984, shows that the chances of a girl surviving the period between one and five years of age are barely half those of boys. In fact, the Karimpur data from the late 1970s and early 1980s bear a remarkable resemblance to the Khanna Study data of the 1950s.

Examining Son Preference

Which girls, then, are most susceptible to childhood death? These two communities have responded differently to modernization. In Ludhiana, where fertility has been drastically reduced and hence fewer children are born, the higher-birth-order daughters (i.e., those with older siblings), especially those born to educated women, are most at risk. In Karimpur, however, the daughters of the poor appear to have become targets of the discrimination that leads to early death.

Strong son preference clearly reveals itself in statements about desired family size. In Ludhiana in 1984, younger women preferred "somewhat fewer than two sons and a half a daughter. Educated women preferred even fewer sons and daughters than this". Older women, when asked how many more children they desired, showed a marked preference for sons, so that educated women between 25 and 29 wanted virtually no daughters - even if they had no living daughters. Karimpur mothers, especially the poor, expressed less of a preference: they often said, "whatever Bhagvan [God] gives, we will have." Others cited the government slogan "two sons and one daughter." This community, where birth control is largely unknown and mortality is still high, appears to lack any preferences for family size.

Ludhiana women's strong son preferences are reflected in two patterns of child mortality. First, "the data on sex differences in child mortality by birth order show a steep rise in mortality of girls at birth order four and higher". Das Gupta goes on to state, "Fourth and higher births appear to be geared toward achieving the desired quota of boys." Hence, there is a particular subset of female children (those at higher birth orders) that is most subject to excess mortality.

Second, the burden of excessive mortality is even more likely to fall on girl, born to mothers who already have one or more surviving daughters.

Among older women (aged 30-59), this subject of daughters has 45 percent higher mortality than their siblings, but this gap increases to 71 percent among the children of younger women (aged 15-29)...The mortality of girls born to mothers who already have one or more surviving daughters is similar (or perhaps greater) among educated mothers as compared with uneducated mothers...It seems that these girls are subjected to increasing concentrations of excess mortality relative to other children if their mothers are younger, and the more so if the mothers are educated.

Among women with one or more years of education, nearly three fourths of women aged 30-34 report ever using contraception, and they have an average of only 2.78 living children. Since fertility levels are falling, younger cohorts of educated women are likely to have an average completed family size of fewer than three living children. Since these women continue to want one to two living sons, they are under greater pressure not to have more than one surviving daughter than is the case for uneducated and older women. who have higher fertility.

At the same time educated women experience substantially lower child mortality than others because they are better able to care for their children. By the same token, they can also keep the mortality of undesired children high by withholding the requisite care. Thus through better control of both fertility an mortality, they are better placed than their uneducated counterparts to mesh their achieved with their desired family size and sex composition.

Unlike Ludhiana mothers, who use lower fertility and sex-specific child mortality to achieve their desired family composition ("somewhat fewer than two sons and half a daughter"), poor women in Karimpur use high fertility and sex-specific child mortality to maximize the number of surviving boy, in attempts to insure family welfare. As in Ludhiana, mortality cannot be understood without also considering fertility behavior and the overall shape of the resulting families. The poor are maximizing male survival. In a society that sees male children as insurance against old age and poverty, the poor find it advantageous to have more living males. Further, the increasing devaluation of female children is due particularly to (1) lost employment opportunities for poor women; (2) increased dowry demands at all social levels; (3) increased lower class male dominance, due to greater employment opportunities for men; and (4) great disparity in male and female education among the poor. The resulting increase in the marginalization of poor women is leading the poor to continue high fertility combined with sex-specific mortality in order to have more sons.

In Karimpur, the ratio of male to female mortality has declined from 84.8 in the period 1932-1951 to 74.5 in the period 1972-1984. That is girls are 10 percent less likely to survive than their brothers in the early 1980s, in contrast to the 1930s and 1940s. This skewing of the male-to-female mortality ratio can be attributed solely to shifts among landless laborers and subsistence farmers: in this community the rich have begun to favor female children. The ratio of male to female mortality among landless laborers and subsistence farmers has declined from 88 during the period 1952-1971 (.84 for the period 1931-1951), to .70 during the period 1972-1984. Meanwhile, the radio of male to female mortality among the rich has increased from .72 in the period 1931-1951, to .84 for the period 1952-1971, to .71 in the period 1972-1984.

Birth rates, in the meantime, have increased. In the period 1969-1975, the crude birth rate for the poor was 41.8 while it was 35.1 for the Brahman landlords. Both Brahman and poor birth rates increased from 1975 to 1983; the poor rose to 47.5, and the Brahman to 40.1. There is no doubt that the poor are having more children now than even before.

Karimpur resident are blatant in their discrimination against female children. A toddler, about to topple off a cot onto a brick courtyard, isn't prevented from falling, as her grandmother comments, "Let her, she is only a girl." A young mother of three daughters induced a late-term abortion, almost killing herself, because she was convinced that the fetus was another girl. The birth of a daughter is not celebrated unless it is the firstborn child - and sometimes not even then. None of the birth songs honor the birth of a daughter. The auspicious head-shaving ritual known as mundan is often eliminated for daughters, especially in poorer families. Daughters are breastfed for shorter periods - which contributes to female malnutrition - because the couple shortens the interval between pregnancies in order to conceive a son. The ratio of female to male deaths from malnutrition is 1.9 - that is, girls are almost twice as likely to die of malnutrition. It is not uncommon to find three-and four-year-old girls so malnourished that they are unable to walk. In the one set of mixed-gender twins that survived infancy, the boy at age one was fat and glowing with health, while his sister was scrawny and had yellowed hair.

In Karimpur, high fertility in conjunction with a general mortality decline increase both the number of daughters who can be "neglected onto death" and the population of sons who might eventually obtain jobs and hence increase family prosperity. In Ludhiana, low fertility in conjunction with sex-specific mortality targeted at higher-birth-order daughters allows families to achieve their desired family composition. Parents in both these communities have greater control of household composition through combining fertility practices with high sex-specific mortality than they would by controlling births alone. Most critically, however, the evidence from these two areas of northern India points to the systemic nature of discrimination against women, a discrimination that puts female life chances in jeopardy perhaps even more now than in the past.

Table 1

Growth of Female Population in India, 1907-1971

Total Male Female Females

population population population per 1,000

Year (millions) (millions) (millions) males

1901 238 121 117 972

1911 252 128 124 964

1921 251 128 123 955

1931 279 143 136 950

1941 319 164 155 945

1951 361 186 175 946

1961 439 226 213 941

1971 548 284 264 930

Source: Govt. of India 1984:10

Table 2

Infant and Child Mortality Rates (deaths per 1,00 live births) by Age at Death, Khanna 1965-1984, Khanna 1957-1959Age at death (months)(a)

-1 1-11 0-11 12-23 24-59 0-59

Khanna, 1965-84

Males 50.7 27.1 77.7 9.4 8.2 95.3

Females 43.0 51.3 94.3 18.5 12.6 125.4

Total 47.0 38.6 85.6 13.8 10.3 109.6

Male/female 1.18 0.53 0.82 0.51 0.65 0.76(b)

Khanna, 1957-59

Total 73.5 82.7 156.2 72.2(c)

Male/female 0.86 0.44(d)

(a) Female mortality is probably higher than estimated here, because the sex ratio at birth indicates that female live births are underreported by approximately 5 percent. Many of these females are likely to have died, perhaps at early ages.

(b) The male/female ratio in the 0-4 year mortality rate in Khanna, 1965-1984, is similar to that for Punjab state, 1971-1975 and 1975-1980 (Dyson 1987).

(c) From Gordon et al. 1965.

(d) Rate calculated per 1,000 population in that age group.

Source: Adapted from Das Gupta 1987:81.

Table 3

Infant and Child Mortality Rates (deaths per 1,000 live birth) by Age at Death over Time in Karimpur

Male Female

mortality mortality Combined M/F

Time period rate rate rate ratio


Died 1-30 days 173 170 171 102

31-365 days 151 178 163 85

1-5 years 100 152 124 66


Died 1-30 days 145 133 139 109

31-365 days 96 166 130 58

1-5 years 102 111 106 92


Died 1-30 days 122 110 116 111

31-635 days 67 112 86 60

1-5 years 35 68 49 51

Table 4

Child Mortality by Birth Order and Sex, Khanna, Based on All Birth to Women Aged 15-49 (deaths 0.4 years per 1,000 live births)

Birth order (not disaggregated by sex)

Sex of child 1 2 3 4+

Male 126.8 96.7 100.9 99.3

Female 95.7 119.1 116.1 152.7

Total 112.1 107.5 108.1 124.5

Male/Female 1.32 0.81 0.87 0.65

Note: The birth order 1 data for males and females do not reveal son preference, but follow the usual pattern. The combined sex figures on mortality by birth order show the typical J-shaped curve: mortality for the first birth is a little higher than that for the second and the third after which it rises again.

Source: Das Gupta 1987:82

Article copyright Cultural Survival, Inc.

Cultural Survival helps Indigenous Peoples around the world defend their lands, languages, and cultures as they deal with issues like the one you’ve just read about.

Learn More

To read about Cultural Survival’s work around the world, click here. To read more articles on the subject use our Search function and explore 40 years of information on Indigenous issues.

Do More

For ways to take action to help Indigenous communities, click here.


We take on governments and multinational corporations—and they always have more resources than we do—but with the help of people like you, we do win. Your contribution is crucial to that effort. Click here to do your part.