What do you think when you hear "6%"?
Chances are, if you're interested in the Civil War or Civil War memory, you may recognize 6% as the number of white Southerners who owned slaves in 1860 (some say 5%. Whatever). In many apologetic texts, this number magically morphs into 6% of Confederate soldiers (or fewer!). Nevermind that recent scholarship puts the number of Confederate soldiers who lived in slaveowning families at well over a third.
This statistic is always dragged out to "prove" that the average Confederate soldier was absolutely not fighting to preserve slavery. It's not just the crazies who argue that very few Confederate soldiers were slaveowners — this stat is pretty well accepted by people who are not inveterate Lost Causers. "[T]he vast bulk of Confederate soldiers did not own a single slave," writes Robert Mackey of the Huffington Post "and many from the most rural areas had never even seen a black man or woman in their lives."
Where did everybody get this number?
It seems that the argument that minuscule numbers of Southerners were slave owners come from sophisticated mathematical formulas that divide the number of slaveowners on the 1860 Census (316,632 in the 11 Confederate states) by the total population of white Southerners (5,447,220 in the CSA). That gets you 5.8%.
I am no math genius, but there seems to be a problem here. The 1860 Slave Schedule lists slaves under the name of the legal owner, who is generally a man and the head of a household. The women and children living in the household were not legal slaveowners, but their inclusion in the data dilutes the percentage of white Southerners who owned slaves to the point of absurdity. Using similar logic, you could "prove" that fewer than half of Americans own cars.
I spent some time looking over the 1860 census data at UVA's wonderful census website and on the library edition of Ancestry.com. Here are some of my preliminary findings:
1. Over half of all white Southerners were under the age of 20.
2. Over 30% of Southern households had slaves.*
*Maybe. I came up with these percentages by dividing the number of slaveowners by the number of households in each state. It is entirely possible that a household could have more than one slaveowner listed in the Slave Schedule, but my preliminary readings of that document indicate that that's pretty rare. These percentages should not be interpreted as a random sample of households, but as rough numbers. My real point here is not that these numbers are perfect, but that 6% is way, way off.
3. The 1860 slave census both overcounts and undercounts slave owners.
The 1860 Slave Schedule overcounts some slaveowners (those with many slaves in more than one county). Yet, it also undercounts slaveowners and those living in slaveowning households. For example, a woman who brings personal attendants with her to her husband's home at marriage is a slaveowner, but she is unlikely to be listed as one on the census. Similarly, the guardians of slaveowning minors (or unrelated women) are often listed as the owners on the census, as in the case of William G. Allen of Clarke Co., Alabama:
4. The 6% figure obscures the number of white Southerners who lived in slaveowning families. Duh.
Let's take the case of William C. Riddle of Washington Co., Georgia. Riddle owned 92 slaves. On the 1860 Slave Schedule, he is the only member of his household listed as an "owner." This household also included Riddle's wife, their five children, and two 26-year-old male tutors.
In 1901, Samuel French wrote, "let it be known that the Confederate army was not an army of slave owners." Confederate apologists near and far have taken that tidbit and run with it. It turns out not to be entirely true. Some will say that the 37% quoted by Joseph Glatthaar in General Lee's Army is still small, but I would point out to them that Glatthaar is only talking about soldiers who "either owned slaves themselves, or the parents or family members with whom they resided in 1860 owned slaves" (Glatthaar, 468). I'm not sure whether he includes those tutors and overseers, who were not family members in their 1860 households. In any event, it is a logical fallacy to conclude that just because someone didn't own slaves he neither aspired to ownership nor supported the slave system. For more on that, see Masters of Small Worlds on the importance of white male mastery over the private household and dependents. The slave system upheld the independence of all white men, whether they owned slaves or not, and all had a vested interest in keeping it up and running.
P.S. Also, I know that this is going to come up, so I'll nip it in the bud. Yes, some women and some free African-Americans owned slaves. Their numbers don't change my conclusions. Since women who are listed as slaveowners are likely to be heads of households, their inclusion doesn't change the percentages listed in part 2. The number of black slaveowners is very small, certainly too small for it to make sense that 2 of the first 4 Google hits for "Confederate slave owner" are dedicated to the subject. From these preliminary forays into the records, I'd say women make up fewer than 10% of slaveowners (that's an educated guess, not a quote-worthy number) and the largest estimate of black slaveowners I've ever seen is 5,000. Maybe for my next project I'll investigate those numbers. After all, there were fewer than 150 black slaveowners in Charleston in 1860. I've seen the number 3,000 quoted for Louisiana, but the slave schedule makes me a little skeptical.
Look at this page from Bernard Co., Louisiana (click to enlarge). There are two names in the "Name of Slave Owner" column that belong to black men — "Big James slave" and "Charles slave." They are both listed as being 100+-year-old fugitives who are listed on the last line along with the total number of slave houses on the plantation. I don't think they were really slaveowners, but that's how they're listed. How were these two men counted by the indexers? It's obvious that there were some black slaveowners, but I wonder about the numbers when the census contains ambiguous entries such as these.