«Volume Title: Strategic Factors in Nineteenth Century American Economic History: A Volume to Honor Robert W. Fogel Volume Author/Editor: Claudia ...»
The probability of sale at different ages for slave children can be estimated by using the exact ages at sale provided by 42 of the 109 ex-slaves who reported being sold. Table 11.6 outlines the computation of the probability of sale for slave children, with the results reported in column 5. Through age sixteen, the slave child faced roughly a 20 percent chance of being sold away from the family. Basing these calculations on the subsample of people who spoke of their family history would raise this probability to about 26 percent.
The table also shows the relatively low probability of sale before age nine.
Through age nine the cumulative probability of sale was just 5 to 7 percent.
From age ten to sixteen the probability increased to 20 to 26 percent.
The probability of sale indicates that a significant number of slave children were sold from their families, a finding that may indicate the tendency of some slaveowners to break up slave families. There are indications, however, that slaveowners did not completely disregard the slave family when making decisions on slave sales. If a slaveowner desired to sell a slave, he had the option of choosing from a group. If the decision were random, we would
20. Rawick, vol. 18, Fisk University Narratives, pp. 226-27.
342 Stephen Crawford
“Derived from the percentage of ex-slaves who reported being sold at that age among all who gave age at sale.
bDerivedby multiplying the percentages in column 1 by 109, the total number of ex-slaves in the entire sample sold from their families.
cDerived by applying the age distribution of the subsample of ex-slaves who gave their exact age, 1,167, to the entire sample, 1,916.
dColumn 3 divided by column 4.
expect the probability of sale to be equal for all slave children. A detailed look at the 109 ex-slaves sold from their families shows that 41 of them, or 38 percent, were subsequently sold at least once. Since the probability of sale within the entire sample was 5 to 7 percent, the probability of being resold once the slave child was initially separated from his or her family increased substantially. Correcting for the age distributions of the different samples would bring these probabilities somewhat closer together, but the fact remains that slaveowners showed a preference for selling slave children already separated from their families.*’ Sale, transfer, and, to some extent, the marriage of the ex-slave are the added information that converts the Family of Origin distribution into the household distribution. The probability that a slave experienced any of these events increased with age. Table 11.7 cross-tabulates this household distribuAn alternative explanation is that disobedient or naughty slaves were sold. If so, it would suggest that young slaves who were separated by sale from their parents did not become effectively socialized into the slave system.
343 The Slave Family
tion by the ex-slave’s age at the end of slavery. Roughly 80 percent of the exslaves were born into either a two-parent consolidated or a two-parent divided-residence family. The proportion of two-parent families steadily decreases as the age of the ex-slave increases. The decrease is due to the breakup of slave marriages through death or sale and the sale or transfer of the slave children. These two factors reduce the percentage of children living in twoparent families to 52 percent by age 12 to 15, and to 40 percent by age 20 and over.
The bottom three rows of Table 1 1.7 show the movement of slave children away from their families into the three special household situations: alone in the quarters or in the master’s house, and married in their own household. The three categories increase steadily across the age distributions from only 1.4 percent who lived separately by age 0 to 3, to 33 percent living separately by age 16 to 29. The incidence of the three household types is best examined by looking at separate household-by-age cross-tabulations for male and female ex-slaves. These cross-tabulations are presented in Tables 11.8 and 11.9.
As the age of the ex-slave increases, the pattern of children leaving the family differs significantly. Female slave children were more likely to be separated from their families and more likely to live in the master’s house than were male slave children. For instance, in the 12-15 age group, 12.5 percent of the males compared with 30 percent of the females were alone either in the quarters or the master’s house. This substantial difference was due to the larger number of females living in the master’s house, with close to 15 percent of females 12 to 15 years of age in the master’s house compared with almost none of the males. The percentage for males and females in the master’s house and quarters does not become the same until the 20-25 age group when roughly 35 percent of both the males and females were apart from their families. The proportion of male slaves in the master’s house increased dramatically for slaves aged 16 to 25. For the female slaves, the proportion in the Stephen Crawford
master’s house peaks among the 12-15 age group and then falls off steadily.
Slave girls who nursed, cooked, and cleaned were useful at a younger age than were slave boys who acted as personal servants and coachmen.
Tables 1 1.8 and 1 1.
9 include information from the narratives regarding the age at which slaves married. The sample is small because few of the ex-slaves interviewed had reached marriageable age during slavery. After all, a slave twenty years old in 1865 would have been a 92-year-old ex-slave informant in
1937. The reported incidence of marriage among the ex-slaves who did reach marriageable age was low in the narratives. This might reflect the lack of interviewer interest in the subject or the ex-slave’s focus on initial family life.
Even if the absolute levels are suspect, the movement in the percentagemarried at different ages provides information on the age at which males and females tended to get married.
Female slaves were married at a significantly earlier age than males, although very few females were married before fifteen years of age. The first large group of female slaves to marry was in the 16-19 age group where 16 345 The Slave Family percent reported being married. The percentage married doubles between the 16-19 and the 20-and-older age groups. Slave girls began marrying after age 16, but the largest percentage waited until they were 20 or older. The first jump in the percentage of male slaves married does not come until the 20-andolder category. In fact, the small number who said they were married suggests that many male slaves probably waited until their mid to late twenties before getting married.
The household information together with the previously discussed estimates of sale can be used to provide a view of the permanence of slave families. Roughly 75 to 80 percent of the slaves in the narratives were born into two-parent (consolidated and divided-residence) families. By age nineteen, close to 50 percent of the slaves were still members of such families. Thus roughly 40 percent of the slave children born into two-parent families experienced the loss of a parent by death or sale or were themselves sold or transferred from the family. Roughly 20 percent of slave children never experienced life in a two-parent household-because they had a white father or a slave father whom they never knew, their family was never fully formed. To grossly simplify the slave family structure, 80 percent of the children were born into two-parent households and 40 percent of these would experience a disruption from death, sale, or transfer, by age twenty. Twenty percent of the slave children never experienced life in a two-parent family.
The analysis of the family thus far groups all ex-slaves together regardless of location, plantation size, or job of their parents. By cross-tabulating either the family of origin or household distribution with these important factors, subgroups in the slave population can be examined in detail.
Arguments about the families of fieldhands and houseservants have existed since the pioneering work of W. E. B. Du Bois. Du Bois thought that it was only among the houseservants that a strong family existed. Although it is difficult to determine whether Du Bois believed that the relationships within families were weak or that no nuclear family structure existed among fieldhands, his statement implies different family distributions among houseservants and fieldhands. Table 11.10 presents a cross-tabulation of family type
by the job of the ex-slave’s mother (given the low percentage of male-headed families, a comparable analysis is not possible for the job of the father). Quite clearly there was no difference between the family structure of houseservants and fieldhands. It is possible that the family relationship differed in other ways, but there was no significant difference in the proportion of two-parent and one-parent families.
The two most important factors affecting the household distribution were the size and location of the ex-slave’s plantation. The plantation-size crosstabulation is presented in Table 11.11. The percentage of children in twoparent, consolidated households was lower on farms with one to fifteen slaves than on those with sixteen or more slaves. The lower percentage is offset in part by the higher incidence of two-parent divided-residence households on small farms. Taking all of the two-parent households together, however, only 35 percent of the children on farms with one to five slaves and 52 percent of those on farms with six to fifteen slaves were in two-parent households. On slaveholdings of sixteen or more slaves, 67 to 7 3 percent were in two-parent households. The smaller slaveholdings had a higher incidence of one-parent households and of children separated from their parents. Twenty-eight to thirty-five percent of the households on small farms were one-parent compared with only 14 to 19 percent on the larger units. There was also a higher percentage of children living apart from their parents on the smallest farms.
There thus appear to be two household distributions, with the separation coming at roughly fifteen slaves. Below that level, the slave farm may have been too small to provide marriage partners. More likely, small farms had grown or decreased through purchase or sale with resulting breakup of marriages and the separation of children from their parents. On plantations with sixteen or more slaves the two-parent family predominated, although even on these plantations the percentage of slave children living apart from their parents averaged 10 to 20 percent.
After plantation size, it has been assumed that the most important factor affecting the distribution of slave households was location. The story of the movement of the locus of slavery from East to West is well known. This movement separated the South into slave-importing states (Deep South and Southwest) and slave-exporting states (Southeast).
This movement did not, however, create two distinct household distributions. Table 11.12 shows that in both regions, 55 to 60 percent of slave children lived in two-parent consolidated or divided-residence households. The only real difference is the split between these two household types. In the slave-exporting states roughly 15 percent of the slave children grew up in twoparent divided-residence households compared with only 5 percent in the importing regions. The percentages in all other categories are similar.
The differences in the divided-residence household percentages in exporting and importing states could in large part be due to the effect of plantation size. Tables 11.13 and 11.14 present cross-tabulations of household type by plantation size within the two slave regions. Some of the differences already noted in the discussion of plantation size are again present. The proportion of two-parent consolidated households in both regions is much higher on slaveholdings of more than fifteen slaves. Two-parent divided-residence and oneparent residence show significant differences in the two regions. The dividedresidence households were more prevalent in the longer settled exporting states and within the region on small farms. Divided-residence households were comparatively rare in the importing region, owing, at least in part, to the greater geographical distance between slave farms in the new regions.
The higher percentage of female-headed households raises some questions about family formation on small farms in the importing region. These households were the result of fathers unknown to the slave child or, more probably, breakup by sale or transfer. Slaves on small farms in the importing region either migrated with their masters from the exporting states or were purchased
Note: Based on the ex-slaves’ actual living situation at emancipation after the move. The uprooting of a small slave farm was much more likely to break up a slave family because of the higher incidence of divided-residence families on small farms in the slave-exporting regions. As Josephine Howard who grew up in Texas related, “One mornin’ we is all herded up and mammy am cryin’ and say de gwine to Texas, but can’t take papa. He don’t ’long to dem. Dat de lastes’ time we ever seed papa.”22 The second factor affecting the level of one-parent families was the extent to which owners of small farms, especially in the importing states, purchased slaves. The higher level of slave purchase and transfer is suggested by the 25
22. Rawick, vol. 4 (2), Texas Narratives, p. 164. The Slave Family
to 30 percent of slave children on the small farms in importing states who resided alone in the quarters or the master’s house. The tendency to acquire slaves by sale or transfer undoubtedly led to the purchase of slave mothers and some or all of her children. These purchases, along with the breakup of divided-residence families when small farms moved, fueled the growth of one-parent households in the slave-importing region.