I was listening to a podcast recently that talked about how people’s family situations and upbringings affected how they turned out. It struck me that we often talk in this way – e.g. “My dad was an alcoholic and I saw how it hurt our family so I never drank” or “My parents scrimped and saved throughout my childhood so I became very frugal also”.
But how do we know that something was the cause of how we turned out? Maybe you would have abstained from alcohol or become frugal regardless. It’s also very possible that a sibling reacted in a different way to the same family circumstances. Perhaps your brother followed in your father’s footsteps and became an alcoholic also. Or perhaps your sister hated having to scrimp and save throughout childhood and felt liberation in being able to spend freely as soon as she started earning for herself. Or maybe they end up drinking or spending moderately, like most people do.
Most of us know that correlation does not equal causation. We hear people say this and we think, “of course, I know that”. But it’s hard for us to internalise. It’s so tempting to see causation whenever we see two things that feel like they should be somehow related.
The way we speak about causation is also sloppy. We usually just say “A caused B”, without explaining what “caused” means. But “caused” can mean different things.
First off, it’s important to distinguish between probabilistic, sufficient, and necessary causes:
- A probabilistic cause means that if A happens, the chance of B happening increases. For example, smoking causes cancer. That doesn’t mean smoking will always cause cancer. Nor does it mean that you can’t get cancer if you don’t smoke. It just means that smoking increases the chance or probability of you getting cancer.
- A sufficient cause means that if A happens, B must happen. For example, a prolonged lack of oxygen by itself is sufficient to cause brain death. It will always cause brain death.
- A necessary cause means that unless A happens, B cannot happen. For example, humans’ transition to agriculture was a necessary cause of the population explosions that followed. There was no way human societies could’ve supported millions of people from hunting and gathering alone.
Next, we should also recognise that most things also have multiple causes. Each cause is then just a contributory cause. This idea is also known as multivariate causality. For example, a Freakonomics podcast explains that the reason crime went down seems to be explained in part by an increase in the availability of abortion and in part by lead being phased out of gasoline.
Then there’s the different levels of causality (as explained by David Graeber in Bullshit Jobs) or the different between ultimate and proximate causes (as Jared Diamond put it in Guns, Germs and Steel). This is like a young child repeatedly asking “why”? Why can’t we have a dog? Because our landlord won’t allow it. Why do we have a landlord? Because we can’t afford to buy a house just yet. Why can’t we afford to buy a house? Well, that’s probably a lot of reasons that feed into that one.
So causation is actually really tricky but we tend to think of it as being easy. It feels so intuitive to us. We start asking “why” when we’re children, and receive simple explanations that help us make sense of the world. It is also comforting to think that we understand how the world works and why things turn out the way they do. But the world is complex. Causality explanations are often an illusion because objective ignorance is much higher than we typically assume it to be. There’s a big difference between a good story and a full explanation of the causal relationship between two things.
When we hear people say that A “caused” B, or that B happened “as a result of” A, we should make sure we understand what type of causation they’re talking about. People will rarely spell it out.