The Civil War was about slavery. However, it is a common canard among certain strains of Confederate apologists to deny this fact. How might we go about disproving such a claim?

Well, the Civil War’s roots in slavery are widely recognized, both by modern-day historians and the war’s belligerents themselves.

Quick cites:

- Abraham Lincoln’s Second Inaugural: “These slaves constituted a peculiar and powerful interest. All knew that this interest was somehow the cause of the war.”
- South Carolina’s Declaration of Secession: “But an increasing hostility on the part of the non-slaveholding States to the institution of slavery, has led to a disregard of their obligations [under the Constitution] . . . the constituted compact has been deliberately broken and disregarded by the non-slaveholding States, and the consequence follows that South Carolina is released from her obligation.”

But suppose that we were denied the opportunity to use any textual sources on the question. How might we go about demonstrating that the Civil War was, in fact, about slavery?

If words are off-limits, let’s try numbers!

As of the outbreak of the Civil War, the United States had 33 states. (Having already admitted Oregon, but not yet admitted Kansas.) Of these states, 15 were slave states and 18 were free states.

The Confederacy ultimately consisted of 11 states, not counting the putative Confederate governments of Kansas and Missouri. All 11 of those states were slave states.

(As an aside, this provides a really elegant demonstration of why we say the Civil War was “about slavery” rather than “about abolition.” One of the belligerent governments was exclusively slaveholding, while the other was a hybrid regime. We can imagine a counterfactual civil war, where New England seceded so as to no longer be subject to the slave power, which we might more sensibly describe as being “about abolition.”)

That number seems suggestive. But how suggestive is it?

Well, a quick way to tell how unlikely a particular sort of outcome is, is to assume all outcomes in a larger class are equally likely, and ask how many of them belong to the particular subgroup we’re interested in.

So suppose that the 33 US states in existence as of 1860 were definitely going to have a Civil War, in which 11 of them seceded. How likely was it that all 11 would be slave states?

Well, there were 15 slave states, and so there are 15-choose-11 Confederacies that could be formed by picking exclusively from among the slave states, or 1365 all-slave Confederacies.

But there were 33 overall states, and so there were 33-choose-11 Confederacies that could be formed overall, or 193,536,720 possible Confederacies.

So in the realm of pure chance, there was about a 0.00000705 likelihood, a 7 in a million chance, that a random Confederacy would be purely formed of slaveholding states.

Maybe we defined our problem too narrowly? Because of course, the god Mars, Bringer of War, did not appear in the sky in 1860 and shout “okay, eleven of you on one side, twenty two on the other, four years, winner take all.” Instead we could ask: of all possible secessionist groups, what fraction of them would consist exclusively of slave states?

The number of possible subsets of a set is 2 to the power of the number of elements in that set (because every element is either in, or not in, the subset), so there were 2^15, or 32,678 possible Confederacies that were exclusively slaveholding. Technically we should subtract 1, because one of those Confederacies is the null set, with zero elements, and you can’t have a Civil War if everyone is on the same side. So 32,677 possible slaveholding confederacies. Even this is generous; if, say, only Maryland and Virginia seceded, we probably **wouldn’t** say that the war was “about slavery”, because some other issue would be required to pull MD and VA away from every other state—but that’s harder to quantify, so we’ll skip it for now.

On the other hand, there were 2^33, or 8,589,934,592 possible total Confederacies, subsets of the overall United States. Now, maybe we should divide those numbers by 2, because the complimentary sets don’t really represent different scenarios. If the Civil War is Texas vs. Everyone Else, that’s pretty similar to Everyone Else vs. Texas—we should probably just define the “Confederacy” as the smaller one of every pair of complimentary sets, because that’s roughly what it means to secede, rather than to push others out. (We might want to define “smaller” in terms of population/economy/landmass rather than number, but whatever.) So that gives us 4,294,967,296 total Confederacies, from which we subtract 1, giving us four billion, two hundred and ninety four million, nine hundred and sixty seven thousand, two hundred and ninety five Confederacies.

As above, we divide the number of possible slaveholding confederacies by the total number of Confederacies, and we get 0.0000076082, 7.6 out of a million. (I confess that the math surprised me; I expected the figure to be far smaller in the second computation.) Either way you slice it, it was fantastically unlikely that the Confederacy would be an exclusively slaveholding institution—unless slavery were connected with the cause of secession.

A stronger counterargument would be to say that the Confederacy had to be geographically contiguous, or culturally unified, in order to form a single body and that there were only a couple of American regions so unified—so the math doesn’t apply. This argument starts to get more qualitative and vastly more complicated, and so I can’t resolve it here. But the original calculations should let us know that, regardless of what other evidence we accept, it’s remarkably likely that slavery is correlated with the cause of the Civil War.