In your opinion, do you think the homosexual exploitation of the male slave had any lasting implications? Explain your perspective.

In your opinion, do you think the homosexual exploitation of the male slave had any lasting implications? Explain your perspective.

In your opinion, how do you think it effected black families after the emancipation of slaves? Explain your perspective.

This does not have to be long just a quick explanation maybe like 4 to 5 senteces for both questions

attached are the articles that was giving

Destroyed by Slavery? Slavery and African American
Family Formation Following Emancipation
Melinda C. Miller1
Published online: 14 September 2018
# Population Association of America 2018
Abstract This study introduces a new sample that links people and families across
1860, 1880, and 1900 census data to explore the intergenerational impact of slavery on
black families in the United States. Slaveholding—the number of slaves owned by a
single farmer or planter—is used as a proxy for experiences during slavery. Slave
family structures varied systematically with slaveholding sizes. Enslaved children on
smaller holdings were more likely to be members of single-parent or divided families.
On larger holdings, however, children tended to reside in nuclear families. In 1880, a
child whose mother had been on a farm with five slaves was 49 % more likely to live in
a single-parent household than a child whose mother had been on a farm with 15 slaves.
By 1900, slaveholding no longer had an impact. However, children whose parents lived
in single-parent households were themselves more likely to live in single-parent
households and to have been born outside marriage.
Keywords Slavery. Family structure . Inequality . Cherokee Nation . Fertility
It was by destroying the Negro family under slavery that white America broke the
will of the Negro people. Although that will has reasserted itself in our time, it is a
resurgence doomed to frustration unless the viability of the Negro family is
restored. (Moynihan 1965: chapter IV)
Demography (2018) 55:1587–1609
https://doi.org/10.1007/s13524-018-0711-6
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s13524-018-
0711-6) contains supplementary material, which is available to authorized users.
* Melinda C. Miller
millermc@vt.edu
1 Department of Economics, Virginia Tech, Blacksburg, VA 24061, USA
Introduction
In March 1965, Assistant Secretary of Labor Daniel P. Moynihan released The Negro
Family: The Case for National Action (1965). He decried the disorganization and
deterioration of black families, citing evidence of widespread illegitimacy, divorce,
female-headed households, and welfare dependence. The weakness of families and
their destruction under slavery, he argued, was at the center of the “tangle of pathology”
that negatively affected African Americans and “serve[d] to perpetuate the cycle of
poverty and deprivation” (Moynihan 1965: chapter IV). A copy of the report was
leaked to Newsweek magazine, igniting a turbulent debate about the impact of slavery
on black families and the extent to which slave family structures were transmitted
across generations.
Empirically establishing the intergenerational transmission of family structures
developed during slavery is challenging and requires a sufficiently large data set that
links experiences under slavery to post-emancipation family structures. Furthermore,
confounding factors may explain the high rates of single-parent families observed
among African Americans. Most notably, when Moynihan wrote his report, blacks
had lower average levels of income and wealth than whites (Pew 2016; U.S.
Department of Commerce 1993). Thus, widespread poverty could have been driving
his findings instead of slavery.
In this study, I exploit unique policies in the Cherokee Nation to disentangle the
effects of slavery and poverty on families and to construct a data set that follows
enslaved families from 1860 to 1900. The Cherokee Nation, located in what is now the
northeastern corner of Oklahoma, permitted the enslavement of people of African
descent. Because the Cherokee Nation joined the Confederacy during the Civil War,
it was forced to renegotiate its treaties with the United States. The federal government
required the Cherokee Nation to allow its former slaves access to free land. Miller
(2011) demonstrated smaller racial wealth gaps and higher rates of black farm ownership
in the Cherokee Nation than in the South. The Cherokee freedmen also had higher
absolute levels of average wealth than southern freedmen in both 1880 and 1900.
Because the Cherokee freed people experienced the treatment of slavery without the
high rates of poverty that blacks in the United States faced, the variation in land policy
between the Cherokee Nation and the United States allows for examination of the effect
of slavery alone—and not widespread poverty—on black family formation following
emancipation.
To empirically estimate the effect of slavery on subsequent family structures, I use a
data set that links slaves and their descendants among the 1860 U.S. Census Slave
Schedules, 1880 Cherokee Census, and 1900 U.S. Census. By providing observations
about a family’s experience during slavery and two subsequent post-emancipation
observations, this sample can explore how slavery influenced black family formation.
I use information about slaveholding—the number of slaves owned by a single farmer
or planter—to proxy for experiences during slavery. A number of studies have demonstrated
that slave family structures varied systematically with slaveholding (e.g.,
Crawford 1980; Dunaway 2003; Fields 1985; Fogel 1994; Malone 1996; Moneyhon
1999, 2002; Pargas 2008). Slave children in smaller holdings were more likely to be
members of single-parent or divided families. On larger holdings, children tended to be
in nuclear families.
1588 M.C. Miller
I find that the size of slaveholding on which a parent lived in 1860 is negatively
correlated with the probability that a child lived in a single-parent household in 1880.
The result is robust to different metrics of slaveholding and the inclusion of controls for
literacy, income, and wealth levels. The effect is large in size; all else equal, a child
whose mother had been on a farm with five slaves was 49 % more likely to live in a
single-parent household than a child whose mother had been on a farm with 15 slaves.
By 1900, the effect of slavery on family structures had begun to fade. Past
slaveholding experiences had limited effect on a child’s tendency to live in a singleparent
household. However, children of families with higher incomes in 1880 were less
likely to head single-parent households in 1900. The most pronounced effect was found
between 1880 and 1900 family status: a child was approximately 20 % more likely to
live in a single-parent household in 1900 if the household head was also a member of a
single-parent household in 1880.
Slavery, Poverty, and Family Formation
Following the release of the Moynihan Report, black families became a recurring issue
of public discussion. The Reagan administration invoked the image of the “welfare
queen” when claiming that black single mothers were overusing social insurance
programs (O’Connor 1998). Black single mothers were targeted during Clinton-era
debates over welfare reform, and the 1996 Personal Responsibility and Work Opportunity
Reconciliation Act contained several provisions specifically aimed at teenage
mothers (Weinstein 1998). President GeorgeW. Bush championed marriage promotion
policies and funded the African American Healthy Marriage Initiative program. In
2008, Barack Obama delivered a Father’s Day speech decrying that “too many fathers
[are] missing—missing from too many lives and too many homes . . . in the African-
American community” (New York Times 2008).
When Moynihan (1965) first expressed concern about black families, approximately
22 % of black children and 7 % of white children lived in single-parent households.1
By 2012, 55 % of black children lived with only one parent, whereas 21 % of white
children resided with a single parent (Vespa et al. 2013). Although the relatively high
rate of single parenthood in the black community is often thought a modern phenomenon,
Ruggles (1994) found its roots in the decades after emancipation. Drawing from
U.S. Census samples, he found that the proportion of black children living without at
least one parent has been at least twice as high as that of white children since 1880.2 In
that year, 29.9 % of black children lived without at least one parent, compared with
only 12.7 % of white children. The gap remained relatively constant for several
decades.
McDaniel’s (1994) analysis of the 1910 census found that black children were more
than twice as likely to live with only their mothers or without either parent than white
children. Morgan et al. (1993) demonstrated that black households were less likely to
be nuclear and more likely to contain extended families in 1910. Furthermore, black
1 Author’s calculations using data from the U.S. Census Bureau (2017: tables CH-2 and CH-3).
2 Differential mortality rates between white and black parents could account for some of this difference.
Ruggles (1994) adjusted for mortality and found that the gap did not decrease in magnitude.
Destroyed by Slavery? Slavery and African American Family Formation… 1589
children also lived without either parent at a much higher rate. Although some of this
difference was due to differential mortality, black families were more likely to foster
children—that is, children lived with relatives instead of either parent. McDaniel and
Morgan (1996) extended previous work by focusing on potential reasons for the high
rate of fosterage in 1910. Black mothers had higher levels of illiteracy, employment,
and absent spouses than white mothers. After these factors were accounted for, white
mothers were still more than twice as likely to live with all their children than black
mothers.
Preston et al. (1992) found that 1910 census data likely understated the incidence of
nonmarital childbearing. Given marriage patterns and male mortality rates, too many
African American widows were recorded. Marital turnover was more frequent than
suggested by divorce or true widowhood rates. These discrepancies may have reflected
women’s attempts to legitimize children born outside marriage. Preston et al. (1992:2)
concluded that “patterns of marital instability among blacks show greater historical
continuity than is suggested by some current discussions.”
Two dominant explanations exist for the persistently high rates of single-parent
households among black families. First, black families faced widespread poverty. The
black-white income ratio was 1 / 4 in 1867 and 7 / 20 in 1900. Blacks owned 1 / 36 the
property of whites in 1880 and 1 / 23 in 1900 (Higgs 1982, 1977/2008). Poverty can
contribute to high rates of single parenthood and low rates of marriage (Coontz and
Folbre 2002). Frazier (1939) observed that it was through acquisition of property and
education that black men were able to form stable families. Those without property
were more likely to experience illegitimacy and the “tide of family disorganization (p.
485).” Men with lower earnings are less likely to marry and more likely to divorce
(Burgess et al. 2003). Poor women are also less likely to marry (McLaughlin and
Lichter 1997).
Poor women had fewer resources available to them in event of an unplanned,
nonmarital pregnancy. Their families may have been unable to exert pressure to
encourage a “shotgun” (forced) wedding. Cohabitation without marriage was possible
when financial barriers prevented legal formalization. Such couples with children tend
to be less stable than those in state-recognized marriages and are more likely to separate
(Osborne et al. 2007). Families may have been unwilling to encourage marriage even if
able. Franklin (1997) noted that adolescents were particularly important to tenant
farmers. Teen marriage could harm a family’s economic security by removing valuable
labor. Bruce (1889) argued that children could serve as valuable dowry to an unmarried
woman because their labor was transferred to a potential husband upon marriage. Black
families in the South, particularly those with lower incomes, had restricted access to
educational opportunities. Lower levels of education are more likely to be found among
single parents than married parents (McLanahan 2009).
High rates of poverty contributed to the high mortality rates and low life expectancies
that southern blacks faced throughout the nineteenth century. Even children who
had been born into stable nuclear families faced a substantial risk of losing one or both
parents. Although black mortality rates declined during the twentieth century, they
remained high relative to whites. Black children continued to face a higher risk of
losing a parent (Tolnay 1999).
The second explanation for high rates of single parenthood is slavery. Du Bois and
Eaton (1899) and Du Bois (1908) believed that experiences under slavery inflicted
1590 M.C. Miller
lasting harm on black families. Du Bois theorized that two sets of sexual mores had
emerged. Slaves who worked in houses or as artisans prized monogamy, religious rites,
and nuclear families. Comparatively, field hands had very different mores that arose
from “no family life, no meals, no marriages, no decency, only an endless round of
toil . . . . ” (Du Bois 1908:47). Female-headed households and the accompanying high
rates of illegitimacy became their hallmarks. Frazier (1932, 1939) argued that owners
encouraged their slaves to form traditional two-parent families. However, on larger
plantations with weak social control, slaves were less likely to form such unions.
Stampp (1956) and Elkins (1959) built on Du Bois’ work, arguing that slavery was a
repressive institution with a pernicious and lasting impact on black families.
Ignited by the Moynihan Report, a wave of revisionist research claimed that slavery
had not destroyed the black family. Gutman (1976) argued that slaves had developed
their own unique sexual mores but were largely members of two-parent families at
emancipation. Blassingame (1972:77) wrote that “the Southern plantation system was
unique in the NewWorld because it permitted the development of a monogamous slave
family.” Genovese (1976) stated that enslaved women had more household authority
than white women but agreed that slave families were often monogamous and nuclear.
Fogel and Engerman (1995:5) argued that the “family was the basic unit of social
organization under slavery. It was to the economic interest of planters to encourage the
stability of slave families and most of them did so.”
These competing views of slave family structure can be partly reconciled if one
avoids an old problem in the study of slavery: “the failure to take adequate account of
the differences between slave experiences and culture on large and small plantations”
(Fogel 1994:182). Although the dynamics of family formation were influenced by a
multitude of factors—for example, location, period, environment, climate, and economic
conditions—recent scholarship examining slave family structures has found a
general connection between slaveholding size and family structure. Two-parent, nuclear
families were more common on larger slaveholdings, and single-parent or separated
families tended to be more common on smaller holdings.
Crawford (1980) analyzed a sample drawn from the Works Progress Administration
(WPA) Slave Narratives and found two distinctive distributions of family structure. On
farms with 15 or fewer slaves, children were more likely to live in single-parent or noparent
families. Approximately one-third (28 % to 35 %) of households were of the
one-parent type. On larger farms, only 14 % to 19 % were. Crawford found no
relationship between a slave’s working in a house or field and family structure. Several
subsequent studies have supported Crawford’s findings. Moneyhon (1999) focused on
WPA slave narratives of former slaves from Arkansas. Although all slaves from his
largest holding category (50 or more slaves) were members of a two-parent household,
only 48 % of those were in the smallest category (19 and fewer). His analysis of the
1860 Arkansas census schedules suggested that slaves living as families were pervasive
on large plantations (Moneyhon 2002). Malone (1996) gathered data on Louisiana
slave households and found that two-parent families were 16.2 % more common on
larger holdings than on smaller holdings. In her study of Maryland, Fields (1985:25)
highlighted the role of the state’s low median slaveholding in disrupting family life:
“small-scale slaveholding does not seem obviously typecast for the role of villainy.
Nevertheless, much of the suffering incidental to slavery in Maryland resulted, directly
or indirectly, from the small size of slaveholdings.”
Destroyed by Slavery? Slavery and African American Family Formation… 1591
Stevenson’s (1996:227) study of Loudon County, Virginia, revealed the “predominance
of matrifocality and the large percentage of slave women in smaller holdings.”
Pargas (2008) studied two distinctive agricultural regions to identify the effect of farm
size on black families. Fairfax County, Virginia, was characterized by small farmers
who grew grains. In 1860, 84 % of slaves lived on farms of less than 20 slaves.
Georgetown, located on the South Carolina coast, was known for its large rice
plantations. Only 9 % of slaves lived on small farms with 20 or fewer slaves in
1860. The larger plantations of Georgetown predominantly housed coresidential twoparent
families. Nonnuclear family structures were prevalent on the smaller farms of
Fairfax. Dunaway (2003) focused her studies on the smaller farmers of the American
mountain south. She found that two-thirds of children were separated from their
parents, and more than one-half of the slave families had single-parent female heads.
This finding is in stark contrast to arguments of the revisionists. However, when
combined with the large plantation studies of Gutman and Fogel and Engerman, these
studies further support Crawford’s (1980) finding of dual family distributions based on
slaveholding size. Gutman’s finding of relatively stable families relied on the records of
larger plantations, with the smallest in his analysis containing 47 slaves.
Why would the size of a slaveholding influence a slave’s family structure? Fields
(1985:25–26) pointed out that “small holdings divided family members among several
owners, exacerbating the potential trauma of sale and attacking the integrity of family
life even when the question of sale, either within or beyond the state, did not arise.
Husbands and wives might live apart for this reason, seeing each other only when
granted permission.” The likelihood of an appropriate partner being found on a smaller
slaveholding was lower, and slaves on smaller holdings were more likely to marry or
have relations with slaves on another farm (Crawford 1980). In his analysis of slave
marriages using Civil War Pension files, Steckel (1982) found that marriages between
men and women with different owners were more common in counties where the
median slaveholding was lower. Not only did these abroad marriages preclude the
existence of a two-parent family living together in a single household, but permission
from multiple slave owners was required before family members could visit each other.
The stresses inherent in such long-distance relationships may have contributed to an
increased propensity of dissolution or divorce. Moneyhon (1999) found that a majority
of such divided slave families in Arkansas lived on smaller holdings.
Furthermore, either owner could further divide the family geographically by selling
a family member. The risk of sale was likely higher on smaller holdings. Genovese
(1976:10) wrote that slaves wished “to avoid sale to masters of scanty means.” Smaller
farmers were more likely to face economically hard times than larger plantation owners
(Burton 1985). By selling slaves to raise cash, slave owners could have separated
families.
A child’s father could have also been his owner. Women on smaller farms were
much more likely to have children with white men than slave women on larger
plantations. Crawford (1980) found that children on holdings of less than 49 slaves
were twice as likely to report having a white father than children on large holdings.
Moneyhon (1999) calculated that 10 % of slaves raised in nonnuclear families reported
that their fathers were also their owners. Steckel’s (1980) analysis of the 1860 U.S.
Slave Schedules revealed that slave children were more likely to be classified as
“mulatto” on smaller slaveholdings. A child born on a mixed farm that had 10 slaves
1592 M.C. Miller
in a border state was seven times as likely to be “mulatto” than one born on a Deep
South cotton plantation with 75 slaves.
Although family formation under slavery was influenced by the size of slaveholding,
no clear empirical consensus exists as to the intergenerational impact of experiences
during slavery on subsequent generations’ family formation. Studies using modern data
have suggested that single parenthood can be transmitted across generations. Högnäs
and Carlson (2012) used 2002 data to find that being born to unmarried parents
increased a woman’s risk of a nonmarital first birth. The effect remained after they
controlled for parents’ socioeconomic status. Using data from 1968–2003, Kearney and
Levine (2009) found a correlation at the cohort level between women’s likelihood of
early childbearing and that of their mother’s cohort. These results were present, but
weaker, at the individual level. Although these studies suggest that slavery’s impact on
families could have lingering effects, Sacerdote (2005) found evidence that the impact
of slavery on African Americans dissipated within two generations. When comparing
the descendants of free blacks and slaves with respect to literacy, schooling, and
occupation, he found that the free blacks outperformed the former slaves for the first
generation. However, there was partial to complete convergence for the next
generation.
Data
Previous attempts to investigate the influence of slavery on post-emancipation family
structure have been hampered by the lack of a sufficiently large data set that ties slave
experiences to subsequent intergenerational outcomes. To that end, I constructed a data
set linking slave families from the 1860 United States Census Slave Schedules to the
1880 Cherokee Nation Census and the 1900 United States Census. This linkage is
facilitated by information from the Cherokee Freedmen Dawes Cards. Here, I briefly
describe the data sources and linking procedure. See Miller (2015) for additional
details, including images of archival documents and an illustrative example of a linked
family.
Between 1900 and 1907, the United States constructed an official roll of Cherokee
citizens, including freedmen. An enrollment card, commonly called a Dawes Card,
with name, age, sex, familial relationship to others on the card, year of tribal enrollment,
location, former slave owner, mother’s former slave owner, and father’s former
slave owner was created for each family. This card’s unique number was recorded next
to each person’s entry on the 1880 Cherokee Census. This census collected basic
demographic information, including age, race, sex, marital status, literacy, and occupation.
3 Agricultural information was recorded for each family, including total acreage
of a farm, number of dwellings and structures, acreage planted in major crops (e.g.,
cotton, wheat, and corn), yields for each crop, and livestock owned.
3 Freedmen in the Cherokee Nation had their race recorded as “col.,” an abbreviation for “colored.” There was
no distinction made between people of all-African versus mixed-race ancestry; “mulatto” was not a racial
category.
Destroyed by Slavery? Slavery and African American Family Formation… 1593
People on the census can be directly linked to their Dawes Card by matching only on
the enrollment card number. Matching names and/or ages is not required. I digitized
census information for all 1,813 freedmen in the 1880 census and collected Dawes
Cards for a sample of 932 of those who were still living in 1900. I used the name, age,
and family information from the cards to locate each person in the 1900 U.S. Census
and found 789, giving a linkage rate of 84 %. Of those found, 359 were men, and 394
were women; the remaining 35 had no information or illegible information on gender in
the 1880 census. The linkage rate was nearly identical for men and women—at 82 %
and 83 %, respectively—suggesting that the linked census sample is not biased with
respect to gender. For each person located, I digitized the 1900 census information for
that person and household. Data collected include literacy, occupation, unemployment,
homeownership, relationship to the head of household, age, marital status, length of
marriage, number of children born to a woman, and race.
Of those located in 1900, 610 had some slave owner information present on their
Dawes Card: 289 people with an owner’s name, 525 with a mother’s owner, and 433
with a father’s owner. I searched for each owner in the 1860 U.S. Census Slave
Schedule for the Cherokee Nation. After finding an owner, I calculated the size of
owner’s slaveholding in 1860. Table 1 summarizes the slave owner matching rates.
Owners of some type were found for 88 % of people. An individual’s owner was
located more frequently than mother’s or father’s owner. For most people, parents’
owners were older than their owner, and several had died before 1860.
Additional slaveholding information could be imputed. For example, if a child’s
mother’s owner was known, then the mother’s own owner was known even if her card
lacked the name. Through this method, 146 people were assigned additional slaveholding
information. Of the 789 people linked to the 1900 U.S. Census, 683 have some
slaveholding information. Table S1 of the online appendix compares those with
slaveholding information to those without. They are overall remarkably similar, particularly
with respect to family structure: 22.4 % of those without slaveholding information
and 22 % with slaveholding information were members of single-parent
households.
Table 1 Linkage rate between Cherokee Freedmen Dawes Enrollment Cards and 1860 Cherokee Nation
slave schedule
Name of Slave Owner Present
on Dawes Card
(1)
Slave Owner Found
in 1860 Slave Schedules
(2)
%
Found
(3)
Owner 289 233 81
Father’s Owner 433 338 78
Mother’s Owner 525 391 74
Any Owner 610 537 88
Notes: Column 1 contains the number of people located in both the 1800 Cherokee Census and the 1900 U.S.
Census for whom the name of a slave owner was present on the Dawes Enrollment Card. Column 2 contains
the number of those slave owners who were located in the 1860 Slave Schedules for the Cherokee Nation. See
the text and Miller (2015) for more details.
Sources: 1860 Cherokee Nation Slave Schedule and Cherokee Freedmen Dawes Enrollment Cards.
1594 M.C. Miller
Identification Strategy
To assess the impact of family structure under slavery on post-emancipation family
formation, I use the number of slaves on a farm or plantation in 1860 to proxy for
conditions associated with family disruptions during slavery. The size of slaveholding
is negatively correlated with the prevalence of single-parent households. Smaller
slaveholdings tended to have higher rates of single-parent slave families or other forms
of separated families. Proportionally, more two-parent families existed on larger
slaveholdings.
I first use the 1880 Cherokee Census data to analyze the impact of slaveholding size
on 1880 family structure. Second, 1900 U.S. Census data examine any lingering effects
of slaveholding size. I incorporate information on 1880 family structure to test whether
independent of slaveholding size, a child’s chances of becoming a single parent in
adulthood were influenced by residing in a single-parent household. Because these
former slave families in the Cherokee Nation have, on average, higher levels of income
and wealth than former slave families in the South, the effect of slavery in both periods
will largely be isolated from the effect of widespread poverty and tenancy.
Although the results will illuminate the effect of slaveholding on family formation of
former slaves in the Cherokee Nation, whether they can also provide insight into the
effect of slavery on families in the southern United States will require that several
conditions hold. First, a similar relationship between slaveholding and family structure
must have existed in the Cherokee Nation and the southern United States. Second,
slavery in the Cherokee Nation and the South must have been similar with respect to
other factors that could have influenced family structure. Third, following emancipation,
governmental policies in the Cherokee Nation and the United States must have
been similar with respect to family formation.
Slaveholding and Family Structure in the Cherokee Nation
The WPA Slave Narratives and Indian-Pioneer Papers include 23 interviews with
former slaves of Cherokees,4 and 17 contain information about family structure.
Although one must be cautious when drawing inferences from a small sample, the
families described by these former slaves are similar in structure to those found by
Crawford (1980) in his much larger sample of WPA records. Four former slaves (23 %)
stated that their owner was also their father; Crawford had found that 15 % to 25 % of
slaves had white fathers. Nine people lived in two-parent families, and at 52 % of the
sample, this is comparable to the 50 % Crawford (1980) found. Of the remaining four
slaves, one had parents who were married, but lived on different plantations. Three had
parents who had died.
The Dawes Cards provide additional information on the connection between family
structure and slaveholding: 485 people in my sample had cards that contained the
names of both mother’s and father’s owners. Table 2 shows how parental separation
varied by the size of slaveholding on which a mother resided. Two-fifths (38 %) overall
reported that their parents had different owners during slavery, and 43 % of people
4 The Indian-Pioneer Papers (1861–1936) were produced by the WPA and documented experiences in Indian
Territory.
Destroyed by Slavery? Slavery and African American Family Formation… 1595
whose mothers resided on farms with five or fewer slaves had fathers with a different
owner. Only 26 % of those whose mothers were on the largest farms did. Although this
metric applies only if a person knew both parents and had a father who was also a slave,
the results suggest that geographically divided families were more common on smaller
holdings.
Other Aspects of Slavery Life Are Similar
Although the experiences of slaves throughout the United States were not uniform,
evidence suggests commonalities in the conditions of slaves in the Cherokee Nation
and the South. Scholars have concluded that “slavery among the Cherokees was little
different than that in the white South” (Littlefield and Littlefield 1976:19). Slave laws
in the South varied at the state level. However, Cherokee slave laws resembled those of
most southern states. Laws concerning marriage and parentage were similar. Slaves
were forbidden to marry Cherokees or whites, and severe restrictions prohibited
Cherokee citizens from marrying a “person of color” (Cherokee Nation 1868:22). A
mother’s slave status determined that of children. This is a key similarity: it is this law
that classified the children of a slave woman and her owner as slaves (Miles 2000).
Additionally, slaves who travelled away from their owner’s farms were required to have
a permit, and patrols looked for slaves without permits. This pass system, which also
existed in the South, served as a barrier for meetings of slave families living in multiple
locations.5
Slaves in the Cherokee Nation engaged in tasks similar to those in the South, with
distinction between field hands and house servants. Overseers directed the field hands,
employing a gang labor system for larger slaveholdings. On smaller farms, slaves
participated in a variety of duties. They were hired out for manufacturing or other work
when the fields were fallow or if in possession of a particularly valuable skill. This
practice served to separate families both in the South and in the Cherokee Nation. See
Miller (2008: chapter 1) for additional historical background on slavery in the Cherokee
Nation.
Table 2 Slaveholding and parental separation
Mother’s Slaveholding N % Parents in Different Households During Slavery
≤5 203 43
6 to 10 128 38
11 to 15 125 31
16 to 30 36 32
≥30 19 26
Total 485 38
Notes: The table includes all people in the 1880 Cherokee Census whose Dawes Enrollment Cards included
the name of both mother’s slave owner and father’s slave owner.
Sources: 1880 Cherokee Census, Cherokee Freedmen Dawes Enrollment Cards, and 1860 Cherokee Nation
Slave Schedule.
5 Table S2 of the online appendix contains a list of selected Cherokee slave laws.
1596 M.C. Miller
Table 3 contains summary statistics for the 1860 Cherokee slave data and the 5 %
1860 Public Use Microdata Flat Sample of the Slave Population of the United States.
The composition of the slave population is similar in mean age and prevalence of
female slaves, and the two data sources have the same median slave holding. The
Cherokee Nation’s mean is lower, reflecting the lack of the very large slave holdings
that existed in the southern states. The number of slave owners in this upper tail was
small in the South. Only 1.83 % of southern slave owners owned more than 56 slaves,
which was the largest holding in the Cherokee Nation.
Post-Emancipation Policies Are Similar
The land made available to Cherokee freed people provided them with an advantage
over southern freedmen with respect to their economic well-being. Besides land,
however, were the Cherokee Nation’s policies toward freedmen significantly different
from those of the southern states, particularly with respect to family formation? The
evidence suggests that they were not.
The United States had a paucity of social services available to poor black families or
single mothers during the periods in question, and the same is true in the Cherokee
Nation. There is no concern that government welfare or social insurance policies
encouraged female-headed households or illegitimate births in either location. Segregation
existed in the Cherokee Nation as in the South. When testifying before a U.S.
Senate commission,William Boudinot, the executive secretary of the Cherokee Nation,
declared, “It is the policy of the nation that the two races should be separated” (United
States 1886:76). Blacks in the Cherokee Nation could not attend Cherokee and white
schools, and social interaction between the races was limited. Additionally, only one
freedman was elected to office in the Cherokee Nation.
The Cherokee freedmen protested that they were denied access to the vote, were
treated unfairly in jury trials, and had access either to no schools or to poor-quality
schools. There were also complaints that Cherokees perpetrated violence against the
Table 3 Summary statistics from 1860 Cherokee and U.S. slave schedules
Cherokee Nation United States
Sampling Rate 100 5
% Total population enslaved 15 32.27
% Female 50.7 49.69
Mean age 19.15 20.13
Slaveholding Information
Mean 6.57 9.72
Median 5 5
Minimum 1 1
Maximum 56 527
Sources: 1860 Cherokee Nation Slave Schedule and the SlavePUMS sample of the 1860 U.S. Census Slave
Schedules (Menard et al. 2004).
Destroyed by Slavery? Slavery and African American Family Formation… 1597
freedmen. These are all situations that southern freedmen faced, indicating that Cherokee
and southern freedmen faced similar treatment (United States 1886).
Family Formation in 1880
For each person in the 1880 data set, there may be slaveholding information for up to
three slave owners: own, mother’s, and father’s. Not all people have three values.
Restricting analysis to, for example, people who have only mother’s owner available
will exclude potentially valuable information. Thus, I construct a “max holding”
measure to include all people with any slaveholding information and maximize sample
size. If the person’s own information is available, I assign that value. If own information
is unavailable, I next assign the mother’s slave owner information. This choice is in
keeping with evidence of the matrifocal nature of slave families. If neither own nor
mother’s slave information is available, father’s slave owner information is used.
To analyze family structure, I examine two potential outcomes: single-parent families
and female-headed families. Using probit analysis, I estimate the probability that a
child lives in each type of family, conditional on an 1860 slaveholding measure. The
general form of the estimated regression is
ProbðFamily Status ¼ 1Þ ¼ Φ β0 þ β1 ð Slaveholding þ γX þ εÞ:
The dependent variable takes a value of 1 if the child lives in the household type of
interest. The sample is restricted to children aged 18 and younger with the appropriate
slaveholding information. Standard errors are clustered at the family level. To assist in
interpretation, marginal probabilities are reported instead of coefficient estimates.
Single-Parent Households
Of the 952 black children under age 18 in the Cherokee Nation, 247 lived in singleparent
households.6 At 25 %, this rate is higher than the 17.6 % that Ruggles (1994)
found for black children in the United States. This difference shrinks if the percentage
of children living without both parents is considered: 34 % of all children (333
children), compared with 29.9 % in the United States.
The data contain slaveholding information for 661 children. Table 4 calculates the
percentage of children living in single-parent households by slaveholding. Because the
oldest children in my sample were born in 1862, very few have personal owner
information. Therefore, the max holding measure is used. I find an inverse relationship
between size of slaveholding and the percentage of children living in single-parent
households.
Figure 1 plots the cumulative distribution of slaveholding for single-parent and twoparent
households. The single-parent distribution lies above the two-parent distribution,
6 A child is said to live in a single-parent household if he or she lives only with one of his or her parents.
Ruggles (1994) restricted his analysis to children aged 14 years and younger. Of the 872 black children aged
14 years or younger in the Cherokee census, 27 % lived in single-parent households. In unreported results, I
use age 14 instead of age 18 to reestimate the regressions in this article. The results remain qualitatively the
same.
1598 M.C. Miller
suggesting that children of slaves who lived on smaller slaveholdings were more likely
to live in single-parent households.
Table 5 compares sample statistics for the two groups of children. For all measures
of slaveholding, the means are significantly higher for those in two-parent households.
Children in single-parent households lived, on average, in poorer and less-educated
households than those in two-parent households. Their household heads were less
likely to be literate and less likely to own land. Their parents owned fewer acres. The
average total values of crops and livestock were lower for single-parent households.
Table 6 reports the results of four pairs of regressions that estimate the effect of
slaveholding on the probability of a child living with a single parent. Columns 1 and 2
Table 4 Percentage of children aged 18 years and younger in the 1880 Cherokee Census and residing in
single-parent households, by 1860 slaveholding status
Slaveholding N % In Single-Parent Household
≤5 89 34
6 to 10 177 28
11 to 15 125 30
16 to 30 201 20
≥30 69 7
Total 661 25
Notes: The max holding measure for slaveholding is used. See the section Family Formation in 1880 for
details on the construction of max holding.
Sources: 1880 Cherokee Census, Cherokee Freedmen Dawes Enrollment Cards, and 1860 Cherokee Nation
Slave Schedule.
0
.2
.4
.6
.8
1
Cumulative Probability
0 20 40 60
Max Holding
Two-parent households
Single-parent households
Fig. 1 Cumulative distribution of slaveholding by single-parent household status in 1880. The cumulative
distribution of the max holding slaveholding measure is plotted for children in single-parent households
(black) and children in two-parent households (gray) in 1880. Sources: 1880 Cherokee Census, Cherokee
Freedmen Dawes Enrollment Cards, and 1860 Cherokee Nation Slave Schedule
Destroyed by Slavery? Slavery and African American Family Formation… 1599
use slaveholding information for the mother, columns 3 and 4 use slaveholding
information for the father, and columns 5 and 6 use the more encompassing max
holding measure. In each pair, the first regression includes only the slaveholding
measure. I then control for the age of the child. Older children had greater exposure
to events that could have separated their parents, such as divorce, unofficial separation,
or death. A literacy indicator variable takes on a value of 1 if the children’s head of
family is literate.
Because children in single-parent families are from households that, on average, are
poorer, I control for three measures of socioeconomics status. Total acres owned is the
amount of land the family reported in the census. Total income from crops is the total
value of all crops reported, measured in hundreds of 1880 dollars. The value of
livestock owned is the total value of livestock in hundreds of 1880 dollars. Families
with no land, crops, or livestock receive zero values. The estimated coefficients on the
socioeconomic variables do not necessarily identify a causal effect. The children of
widowed women, for example, may have lived in households with lower values of crop
income because their mothers were widowed.
In each regression, the estimated value of the marginal probability for slaveholding
is remarkably stable. Slaveholding for mothers varies from –0.010 to –0.012 and is
significant at the 1 % level. For fathers, the estimated value is –0.009 to –0.010 and is
significant at the 1 % level. Finally, the max slaveholding indicator is smaller in
absolute value than the mother’s and father’s indicator and ranges from –0.006 to
–0.007, while also significant. More measurement error may be present in the max
holding measure, leading to a potential attenuation bias and its estimate being smaller in
Table 5 Sample statistics for children in single-parent households versus those living with both parents in the
Cherokee Nation, 1880
Single-Parent Two-Parent
Mother’s Mean Slaveholding 9.69 15.16
[7.31] [12.61]
Father’s Mean Slaveholding 10.26 17.64
[7.04] [14.25]
Max Holding 12.82 17.65
[8.92] [13.50]
% With Literate Head of Household 44.44 60.00
% With Family Land Ownership 49.9 77.2
Mean Landholding (acres) 24.33 55.82
[88.79] [127.53]
Mean Crop Income (1880 $) 392.47 933.36
[2,476.09] [3,141.95]
Mean Livestock Value (1880 $) 304.38 608.66
[505.614] [1,006.83]
Sources: 1880 Cherokee Census, Cherokee Freedmen Dawes Enrollment Cards, and 1860 Cherokee Nation
Slave Schedule.
Notes: Standard deviations are in brackets.
1600 M.C. Miller
absolute value. Although the estimated effects may seem small, they are for a oneperson
change in slaveholding. For larger differences, the effect becomes more
substantial.
In Fig. 2, I plot the predicted probability and 95 % confidence interval that a child
resided in a single-parent household for each measure of slaveholding when all other
variables are evaluated at their means. With the exception of the livestock value, the
included economic variables are statistically insignificant at conventional levels. Livestock
value has a small impact. An additional $100 of livestock is associated with a
1.9 % to 2.6 % decrease in the probability that a child was in a single-parent family, all
else equal.
The literature concerning slave families finds that the impact of slaveholding
size on family structure changes at approximately 15 slaves. To empirically test
this proposition, I construct a series of indicator variables taking the value of 1 if a
child’s max holding measure was below a threshold size. There were 32 variables
in total—one for each slaveholding size in the sample. In unreported results, I ran
32 regressions to discover the first max holding threshold above which a child was
Table 6 Regressions results on slaveholding and single-parent status in 1880
1 2 3 4 5 6 7 8
Mother’s Slaveholding –0.010** –0.011**
[0.004] [0.003]
Father’s Slaveholding –0.010** –0.009**
[0.003] [0.003]
Max Holding –0.007** –0.007**
[0.002] [0.002]
Max Holding >15 –0.129* –0.132*
[0.063] [0.063]
Age of Child 0.005 0.001 0.002 0.002
[0.004] [0.004] [0.004] [0.004]
1 if Head Literate 0.128 0.108 0.105 0.099
[0.110] [0.102] [0.099] [0.099]
Total Acres Owned –0.006 –0.002 –0.005 –0.004
[0.003] [0.003] [0.003] [0.003]
Total Income From Crops 0.018 0.001 0.015 0.015
[0.014] [0.014] [0.012] [0.013]
Value of Livestock Owned –0.019† –0.026* –0.020* –0.020*
[0.011] [0.011] [0.010] [0.010]
Number of Observations 586 586 546 546 661 661 661 661
Pseudo-R2 .0454 .131 .0659 .176 .0278 .101 .02 .0888
Notes: The dependent variable is 1 if a child lives in a single-parent family. Standard errors are clustered at the
family level and reported in brackets. Marginal effects are reported.
Sources: 1880 Cherokee Census, Cherokee Freedmen Dawes Enrollment Cards, and 1860 Cherokee Nation
Slave Schedule.
†p < .10; *p < .05; **p < .01
Destroyed by Slavery? Slavery and African American Family Formation… 1601
statistically significantly less likely to be from a single-parent household. This
threshold occurred at 15 slaves, and the threshold coefficients remained significant
for all greater values. The last two columns in Table 6 report results with the 15
slaves indicator. Children with a max holding value greater than 15 were 12.9 %
less likely to live in a single-parent household. This result is statistically significant
at the 5 % level. When the additional controls are included, the probability
remains significant and rises to 13.2 %.
I consider two additional specifications to confirm the robustness of the results.
Children aged 15–18 were born during the Civil War and may have been born as
slaves. I exclude these children and restrict my analysis to children aged 14 years and
younger.
In Table S3 of the online appendix, I calculate the percentage who lived in
single-parent households by slaveholding. The result is consistent with that for
children 18 and younger: as parent’s slaveholding increased, fewer children
resided in single-parent households. I repeat the regression estimation with the
age-restricted sample. Results are in the online appendix, Table S5. The estimated
coefficients on slaveholding are identical in significance level and magnitude. The
only exception occurs in the fully controlled specification for fathers. The coefficient
increases slightly in magnitude from –0.009 to –0.010. As an alternative
check, I include all children aged 18 years and younger but add an age-
0
.2
.4
.6
Probability Single-Parent Family
0 20 40 60
Slaveholding of Father
0
.2
.4
.6
Probability Single-Parent Family
0 20 40 60
Slaveholding of Mother
0
.2
.4
.6
Probability Single-Parent Family
0 20 40 60
Max Holding
0
.1
.2
.3
.4
Probability Female-Headed Family
0 20 40 60
Slaveholding of Father
0
.1
.2
.3
.4
.5
Probability Female-Headed Family
0 20 40 60
Slaveholding of Mother
0
.1
.2
.3
.4
Probability Female-Headed Family
0 20 40 60
Max Holding
Fig. 2 Predicted 1880 probability of residence in household type, by slaveholding, and 95 % confidence
intervals. The black line in each figure plots the predicted probability of a child being a resident in a singleparent
family (top row) or a female-headed family (bottom row) by slaveholding measure when all other
variables are evaluated at their means. The 95 % confidence interval is plotted in gray
1602 M.C. Miller
slaveholding interaction term to confirm that the impact of slaveholding does not
vary by age. Results are shown in the online appendix, Table S6. The coefficient
on the interaction term is 0 and insignificant in every specification.
Female-Headed Households
Because the public debate over the structure of black families has frequently focused on
the prevalence of female heads, I separately consider the case of children who live in
such households. Indeed, 71.2 % of the single parents in the sample were women.
Table 7 reports the result of regressions with the dependent variable equal to 1 if a child
lives in a female-headed household. The coefficient estimates on the slaveholding terms
are always negative and significant and consistent after the inclusion of additional
covariates. Although the estimates again appear small, the effect of slaveholding on the
probability that a child lived in a female-headed home is economically significant when
Table 7 Regressions results on slaveholding and female-headed households in 1880
1 2 3 4 5 6 7 8
Mother’s Slaveholding –0.009* –0.009**
[0.003] [0.003]
Father’s Slaveholding –0.008** –0.007**
[0.002] [0.002]
Max Holding –0.005* –0.006**
[0.002] [0.002]
Max Holding >15 –0.101† –0.114*
[0.056] [0.051]
Age of Child 0 –0.003 –0.001 –0.002
[0.003] [0.003] [0.003] [0.003]
1 if Head Literate 0.015 0.03 0.023 0.017
[0.077] [0.071] [0.073] [0.071]
Total Acres Owned –0.010** –0.007** –0.008** –0.009**
[0.002] [0.002] [0.002] [0.002]
Crop Income 0.035** 0.022* 0.030** 0.032**
[0.009] [0.009] [0.009] [0.009]
Value of Livestock –0.009 –0.011 –0.012 –0.011
[0.007] [0.007] [0.008] [0.008]
Number of
Observations
586 586 546 546 661 661 661 661
Pseudo-R2 .0441 .172 .0556 .201 .0228 .13 .017 .123
Notes: The dependent variable is 1 if a child lives in a female-headed household in 1880. Standard errors are
clustered at the family level and reported in brackets. Marginal effects are reported.
Sources: 1880 Cherokee Census, Cherokee Freedmen Dawes Enrollment Cards, and 1860 Cherokee Nation
Slave Schedule.
†p < .10; *p < .05; **p < .01
Destroyed by Slavery? Slavery and African American Family Formation… 1603
considered over larger changes in slaveholding size. Figure 2 plots the predicted
probabilities and 95 % confidence intervals.
Family Formation in 1900
Fewer Cherokee freedmen children, both in absolute numbers and by percentage, lived
in in single-parent or no-parent homes in 1900. There were 1,083 children under age 18
in the 1900 sample. Nearly 200 (197; 18.2 %) lived in households headed by a single
person, and 23.6 % lived without one or both of their parents. Although Ruggles (1994)
did not provide estimates for 1900, he estimated in 1910 that 17.1 % of black children
lived with only one parent.
Using probit analysis, I examine two outcomes of interest: single-parent households
and possible nonmarital births. The general form of the estimated regression is
ProbðHousehold Status ¼ 1Þ ¼ Φ β0 þ β1 ð Slaveholding þ γX þ εÞ:
A household with a head, a spouse, and young children has up to six potential values
for slaveholding: mother’s owner, maternal grandmother’s owner, maternal grandfather’s
owner, father’s owner, paternal grandmother’s owner, and paternal grandfather’s
owner. To make the analysis more tractable, I focus on the head’s slaveholding
information, thereby excluding stepchildren of the household head from the analysis.
I again construct a max holding measure. The variable takes the value of the head’s
owner’s slaveholding value. If that is unavailable, the head’s mother’s value will be
used. The head’s father’s information is lastly used.
Single-Parent Households
Panel A of Table 8 summarizes regression results, with the dependent variable taking a
value of 1 if a child lived in single-parent household. For each slaveholding measure,
the results of three regressions are reported. The first has the slaveholding measure as
its only covariate. An indicator variable equal to 1 if the household head was a member
of a single-parent household in 1880 is next included. Finally, the values of crops
grown and livestock owned reported by the head’s family in the 1880 Cherokee Census
are added.
The first row of panel A, Table 8, reports whether the slaveholding measure was
significant at the 5 % level or greater. The coefficient values are provided for statistically
significant results.7 The additional covariates are summarized in a similar manner.
In stark contrast to the 1880 results, the slaveholding measures are almost universally
statistically insignificant. None of the baseline specifications find a statistically significant
relationship between the probability that a child was a member of a single-parent
household and 1860 slaveholding. The only significant variable is 1880 crop income
when size of slaveholding on which the head’s mother resided is the dependent
variable, suggesting that the risk of becoming a single parent was higher for women
from poorer families—a finding that is consistent with the modern literature on single
7 Complete results for this table are available in the online appendix, Tables S6–S9.
1604 M.C. Miller
Table 8 Regressions results on slaveholding and family formation in 1900
Head’s Own
Slaveholding
Head’s Mother’s
Slaveholding
Head’s Father’s
Slaveholding
Head’s Max Holding
1 2 3 4 5 6 7 8 9 10 11 12 13 14
A. Dependent Variable = 1 if Child Lived in a Single-Parent Household in 1900
Slaveholdingmeasure N N N N N N N N N N Y Y
Coefficient 0.006 0.006
Member of single-parent household in 1880 N N N N N N Y Y Y Y
Coefficient 0.216 0.203 0.182 0.201
Total income from crops in 1880 N Y N Y Y
Coefficient –0.008 –0.008 –0.008
Value of livestock owned in 1880 N N N N N
B. Dependent Variable = 1 if Child Was a Nonmarital Birth in 1900
Slaveholdingmeasure N N N N N N N N N N N N
Coefficient
Member of single-parent household in 1880 N N Y Y N N Y Y Y Y
Coefficient 0.202 0.198 0.206 0.202 0.187 0.198
Total income from crops in 1880 Y Y Y Y Y
Coefficient –0.008 –0.009 –0.006 –0.009 –0.008
Value of livestock owned in 1880 N N N N
Number of Observations 335 335 335 455 455 455

The Sexual Abuse of Black Men
under American Slavery
THOMAS A. FOSTER
DePaul University
I N 1787 AN ENSLAVED MAN IN Maryland raped a free black woman.
The story comes to us from the female victim in the incident, Elizabeth
Amwood. One white man, William Holland, had her “Pull up her Close
and Lie Down he then Called a Negrow Man Slave” “and ordered him to
pull Down his Britches and gitt upon the said Amwood and to bee grate
with her.” A fourth individual in this horrific scene, a white man named
John Pettigrew, operating with Holland, pointed a pistol at the unnamed
enslaved man and Elizabeth Amwood. All the while, Holland taunted them
both, asking if it “was in” and “if it was sweet.” Afterward, William “went
up into the Company and Called for Water to wash his hand, saying he had
bin putting a Mare to a horse.”‘
Scholars have suggested that rape can serve as a metaphor for enslavement—
thus applying to both men and women who were enslaved. As
Aliyah I. Abdur-Rahman argues, “The vulnerability of all enslaved black
persons to nearly every conceivable violation produced a collective ‘raped’
subjectivity.”^ The standard scholarly interpretation of how slavery affected
black manhood is perhaps best captured by the comments of one former
slave, Lewis Clarke, who declared that a slave “can’t be a man” because he
I would like to thank all those who helped with the development ofthis article, including
Ramón A. Gutiérrez, Mathew Kuefler, and the participants in the history ofthe intersection
of race and sexuality conference hosted by the Center for the Study of Race and Ethnicity
at the University of California, San Diego. I would also like to thank Estelle Freedman and
Margaret Storey for feedback on earlier versions ofthis article.
‘ Petitions for William Holland, March 1787, Governor and Council, Pardon Papers, box
4, folder 47, Maryland State Archives (hereafter MdSA). See also Sharon Block, Rape and
Sexual Power in Early America (Chapel Hill: University of North Carolina Press, 2006), 85.
^ Aliyah I. Abdur-Rahman argues: “More than simply a condition of black women’s experience
under slavery, rape serves as a useful paradigm for assessing and describing the position
and experience of black people in total under slavery’s brutal regime” (“‘The Strangest Freaks
of Despotism’: Queer Sexuality in Antebellum African American Slave Narratives,” African
American Review 40, no. 2 [2006]: 230-31).
Journal ofthe History of Sexuality, Vol. 20, No. 3, September 2011
© 2011 by the University ofTexas Press, PO Box 7819, Austin, TX 78713-7819
445
446 THOMAS A. FOSTER
could not protect his female kin from being sexually assaulted by owners and
overseers.’ Clark’s concern, the rape and sexual assault of black women and
girls, has been well documented by the historical record. Thelma Jennings
and others have analyzed the literal sexual assault of enslaved women in a
range of contexts.* Physical sexual abuse of women and girls under slavery
ranged from acts of punishment to expressions of desire and from forms
of forced reproduction to systems of concubinage. Slavery violated the
masculinity of black men who were denied the ability to protect vulnerable
female dependents. According to Deborah Gray White, “Those who tried to
protect their spouses were themselves abused.”^ The emasculating psychic
toll. White further argued, could have led men to eschew monogamy or
resist marriage altogether.*
The rape of Elizabeth Amwood reveals that black manhood under slavery
was also violated in other ways that are less easily spoken of (then and now),
namely, the sexual exploitation of enslaved men.” The historical sexual assault
of men and boys is well known, if mostly unarticulated.* The scholarship
on early America shows us numerous instances of rape and sexual assault of
men and boys. Ramón Gutiérrez has argued that individuals of the Native
American third sex, or berdaches, were frequently prisoners of war used for
‘ Lewis Clarke, “Leaves from a Slave’s Journal of Life,” National Anti-Slavery Standard,
20 and 27 October 1842, reprinted in Slave Testimony: Two Centuries of Letters, Speeehes,
Interviews, and Autobiographies, ed. John W. Blassingame (Baton Rouge: Louisiana State
University Press, 1977), 156-58.
* Thelma Jennings, “‘Us Colored Women Had to Go Through a Plenty’: Sexual Exploitation
of African-American Sh.\cV^omcn,^ Journal of Women’s History \, no. 3 (1990): 45-74.
* Deborah Gray White, Ar’n’t I a Woman? Female Slaves in the Plantation South (1985;
New York: Norton, 1999), 146. See also, for example, Daina Ramey Berry, “Swin^ the Sickle
for the Harvest Is Ripe”: Gender and Slavery in Antebellum Georgia (Chicago: University of
Illinois Press, 2007), 81.
‘ White, Ar’n’t I a Woman?, 147.
‘ The rape of adult men in conflict and war today, for example, goes underreported and is
much less discussed than rape of women as a weapon of war. In 2005, for example, in eastern
Congo and northern Uganda, rape of boys and men was a notable feature of the conflicts
there. This has been documented by the United Nations, Human Rights Watch, and Amnesty
International. See, for example, its 2005 Annual Report for Congo (Dem. Rep. of), available
online at http://www.amnestyusa.org/annualreport.php?id=ar&yr=2005&c=COD (accessed
1 August 2010). Also muted in discussions of the rape and sexual assault of men is prison
rape. See, for example. Human Rights Watch, “No Escape: Male Rape in U.S. Prisons,” 2001,
available online at http://www.hrw.org/legacy/reports/2001/prison/report.html (accessed
1 August 2010). On the history of sex and sexual abuse behind bars, see, for example, Regina
Kunzel, Criminal Intimaey: Prison and the Uneven History of Modern American Sexuality
(Chicago: University of Chicago Press, 2008). More often than not, rape of men in prison is
the subject of derisive humor in popular culture masking a deep discomfort and homophobia.
But dismissive humor also hides the deeper threat that male rape exposes—the penetrability
and vulnerability of men.
* Michael Scarce, Male on Male Rape: The Hidden Toll of Stigma and Shame (New York:
Plenum, 1997).
Sexual Abuse of Black Men 447
sex and emasculated. We also know through the handfial of extant sodomy
cases that males have been so abused. The seventeenth-century Connecticut
gendeman Nicholas Sension, for example, sexually preyed on his male
servants. Virtually all of the cases of sodomy that came to the courts in
early America involved individuals violating status boundaries—instructors
on students, masters on servants. None involved peers.’
In the context of slavery, literary scholars have shown that sexual abuse
of men was part of the Spanish slave system in Cuba. Robert Richmond
Ellis argues that the account of former slave Juan Francisco Manzano “has
commonly been regarded as a searing indictment of a physical mistreatment
of slaves” but “can also be read as silent testimony to a kind of abuse largely
unacknowledged by historians of slavery and critics of slave narratives: the
sexual violation of male slaves.” As Ellis points out, the topic has largely gone
unexplored for a wide variety of reasons, including the obvious barrier ofthe
historical record in that “male victims of slave rape left behind no biological
record in the form of offspring” as well as the prevalent homophobia
in traditional Latin American societies, which would have prevented men
from telling their stories given that “male sexual passivity . . . was particularly
stigmatized insofar as it was seen as entailing a loss of masculinity.”‘”
This article uses a wide range of sources on slavery—early American
newspapers, court records, slave owners’ journals, abolitionist literature,
and the testimony of former slaves collected in autobiographies and in interviews—
to argue that enslaved black men were sexually assaulted by both
white men and white women. It finds that sexual assault of enslaved men
took a wide variety of forms, including outright physical penetrative assault,
forced reproduction, sexual coercion and manipulation, and psychic abuse.
It is difficult to determine with certainty the prevalence of the sexual
abuse of male slaves. Martha Hodes concludes that the sexual “coercion” of
black men in antebellum America “lurked as a possibility regardless of how
frequently it came to pass.”” Antislavery movements documented relatively
more instances of such abuse than did previous eras. Nineteenth-century
sources discussing slavery in the South, for example, are more abundant
given the abolitionist movement, which drew attention to sexual depravity
to argue for the immorality of slavery as an institution. Given the variety of
social and cultural barriers to documenting the sexual abuse of enslaved black
men, however, it would be an error to assume that the pattern of surviving
‘ See Thomas A. Foster, ed., Lon¿ before Stonewall: Histories of Same-Sex Sexuality in Early
America (New York: New York University Press, 2007), introduction and essays by Ramón
Gutiérrez and Richard Godbeer.
‘° Robert Richmond Ellis, “Reading through the Veil of Juan Francisco Manzano: From
Homoerotic Violence to the Dream of a Homoracial Bond,” PMLA 113, no. 3 (1998):
422-35.
” Martha Hodes, White Women, Black Men: Illicit Sex in the Nineteenth-Century South
(New Haven, CT: Yale University Press, 1997), 139.
448 THOMAS A. FOSTER
sources reflects the historical practice of abuse. Indeed, the unlikelihood
that cases would have been documented at all suggests that it would be
safe to say that, regardless of location and time period, no enslaved man
would have been safe from the threat of sexual abuse.
In many ways the cases discussed here have been hidden in plain sight.
This article revisits instances that other scholars have cited in their studies
of sex and slavery. In particular, it is indebted to Martha Hodes’s research
on antebellum sex between white women and enslaved black men, Thelma
Jennings’s work on the sexual exploitation of enslaved women, and Deborah
Gray White’s study of plantation life for enslaved women.’^ Through their
painstaking research in slave records, these and other scholars show that
the sexual abuse of enslaved women was ubiquitous. Establishing this now
widely accepted conclusion was itself a challenge, for historians had to argue
against deep-rooted racist depictions of enslaved women as hypersexual.
Turning to the sexual abuse and exploitation of enslaved men builds upon
the perspective ofthis recent literature to challenge the view of black men
as hypersexual and white women as passive and asexual.
Although scholars have acknowledged the sexual assault of enslaved
women, none to my knowledge have highlighted the sexual abuse of enslaved
men. In part, we have taken our cues from the nineteenth-century
abolitionist writers who drew upon certain gender-, race-, and class-infiised
understandings of sexual assault to appeal to a particular audience. As Martha
Hodes reminds us, though, it was not simply that sex between black men
and white women was uninteresting to abolitionists. Individuals recognized
that it was “dangerous to the cause” to insult the virtue of southern white
womanhood.’^ The rape of slave men has also gone unacknowledged because
ofthe current and historical tendency to define rape along gendered lines,
making both victims and perpetrators reluctant to discuss male rape. The
sexual assault of men dangerously points out cracks in the marble base of
patriarchy that asserts men as penetrators in opposition to the penetrable,
whether homosexuals, children, or adult women. This article, therefore,
confronts our own raced, classed, and gendered perceptions of rape and
argues that we have a moral imperative to recognize the coerced sexuality
of enslaved men as rape. Narrowly defining sexual assault along gendered
lines has obscured our ability to recognize the climate of terror and the
physical and mental sexual abuse that enslaved black men also endured.
BLACK MALE BODIES
The sexual exploitation of enslaved black men took place within a cultural
context that fixated on black male bodies with both desire and horror. Sexual
‘^ Ibid.; Jennings, “‘Us Colored Women'”; White, Ar’n’t I a Woman?
” Hodes, White Women, Black Men, 128.
Sexual Abuse of Black Men 449
assault took a wide variety of forms, but the common factor in all was the legal
ownership that enabled control ofthe enslaved body. Winthrop Jordan notes
the conflicting messages embraced by Anglo-American culture as it sought
to control and circumscribe the bodies of enslaved men and women, on the
one hand voicing repulsion for Africans, framing them as beastly, ugly, and
unappealing, while on the other hand viewing them as hypersexual. Anglo-
American culture had a long-standing view of black men as “particularly
virile, promiscuous, and lusty.”‘* Although this view is consistendy ftamed
as a negative one, given Anglo-American cultural norms of moderation
and self-control, it is clear that early Americans also saw erotic possibilities
and beauty in black bodies. We know, for example, that some slave masters
fetishized and objectified women of color, understanding that sexual abuse
was about power and not simply expressing sexual desire. The presence of
antebellum “fetish” markets of light-skinned enslaved women, in particular,
has been well documented by scholars. Edward Baptist, for example, argues
that the antebellum domestic slave trade might be reconsidered as a “complex
of inseparable fetishisms” given the slave traders’ “ftequent discussions
ofthe rape of light-skinned enslaved women, or ‘fancy maids,'” and “their
own relentlessly sexualized vision ofthe trade.”‘^
The evidence also leads us to speculate that an unusual interest in lightskinned
men may have paralleled the more formalized and documented
fetish market in “fancy maids” that Edward Baptist has analyzed. Such an
interest is found in testimony presented to the American Freedmen’s Inquiry
Commission (AFIC), which was established by the secretary of war
in 1863 to document the conditions of those freed by the Emancipation
Proclamation. White abolitionist Richard J. Hinton, for example, testified
that “I have never yet found a bright looking colored man” “who has not
told me of instances where he has been compelled, either by his mistress,
or by white women ofthe same class, to have connection with them.”‘* In
another case, a man testified that a man who had been “brought up in the
family” was also coerced into sex by his mistress, his family connection suggesting
that he was mulatto. We also have some evidence of light-skinned
black men as sexually prized. Testimony to the AFIC included reference to
light-skinned men as “fine looking.”‘^ One man told the AFIC: “It was an
extremely common thing among all the handsome mulattoes at the South
to have connection with the white women.”‘* In the antebellum divorce
‘•’ Winthrop D. Jordan, White over Black: American Attitudes toward the Ne¿iro, 1550-1812
(Chapel Hill: University of North Carolina Press, 1968), 151.
‘* Edward E. Baptist, “‘Cufify,’ ‘Fancy Maids,’ and ‘One-Eyed Men’: Rape, Commodification,
and the Domestic Slave Trade in the United States,” American Historical Review 106,
no. 5 (2001): 1620.
” Richard Hinton testimony, quoted in Hodes, White Women, Black Men, 130-31.
” Ibid.
” James Redpath testimony, quoted in ibid., 127.
450 THOMAS A. FOSTER
case of one white couple, Dorothea and Lewis Bourne, Dorothea’s chosen
lover, an enslaved man named Edmond, is described in the records by more
than one neighbor as “so bright in his colour, a stranger would take him
for a white man.”” Such testimony raises the possibility that in this patriarchal
society the sexual abuse of “nearly white” men could enable white
women to enact radical fantasies of domination over white men with the
knowledge that their victim’s body was legally black and enslaved, subject
to the women’s control.
Although we have no evidence for a sexual fetish market in black male
flesh, historical scholarship shows us that black male bodies might well be
eroticized by white observers. Jordan claims that Anglo-American culture
long held a fascination with the penises of black men and projected both
desire and jealousy upon an objectified and disembodied black phallus.^”
Colonial accounts abound with recorded instances of masters and others
commenting not only on the nudity of slaves but on their bodies with a
certain fascination. As Philip Morgan reminds us, “daily encounters had a
sexual dimension” in part because slaves “wore little or no clothing.” One
observer in 1781 named William Feltman remarked on the reaction this
nudity might provoke among Virginia women, given that “young boys of
about Fourteen and Fifteen years Old” were “virtually naked.” Feltman
quipped: “I can Assure you It would Surprize a person to see these d d
[damned] black boys how well they are hung.”^’ Numerous abolitionist
images also fixate on the black male body as perfection, highlighting
muscular bodies and, in almost pornographic detail, exposed buttocks,
enduring unjust abuse and degradation. William Benemann and others
maintain that the image of whipping exposed male flesh carried a homoerotic
charge—one that mirrored the nearly obscene fixation on whipping
nude enslaved women, as has been suggested by scholars such as Colette
Colligan.’^^ John Saillant’s work on the eroticization ofthe black male body
in early abolitionist literature also contributes to this view that whites found
sexual appeal in black male bodies. He notes that this literature idealized
black male bodies in a manner that included an unusual focus on height,
musculature, and skin color. Accounts in late eighteenth-century and early
” Lewis Bourne divorce petition, Louisa County, Virginia, 20 January 1825, quoted
in American Sexual Histories, ed. Elizabeth Reis (Maiden, MA: Blackwell, 2001), 166-67.
^° Jordan, White over Blaek, 34-35.
^’ Military journal of Lt. William Feltman, 22 June 1781, HSP, quoted in Slave Counterpoint:
Black Culture in the Eighteenth-Century Chesapeake ani Lowcountry, by Philip D.
Morgan (Chapel Hill: University of North Carolina Press, 1998), 398-99.
^^ William Benemann, Male-Male Intimacy in Early America: Beyond Romantic Friendships
(New York: Harrington Park, 2006), 68-69; Colette Colligan, “Anti