non-dominating force

One way to model real-world causality is a bunch of forces working with and against each other. In this view, no individual force determines the outcome. Individual causal effects are just terms in the sum.

Suppose that smoking tobacco gave you magical healing powers, but also still damaged your lungs. Then 'smoking causes lung cancer' would be a term in the sum---there is a causal pathway by which this happens---but it would not necessarily represent the total effect of smoking on your cancer outcomes.

This trips people up in political arguments. Someone will argue that there is clearly a causal effect between policy A and outcome B, and they can even trace the mechanism. But tracing the mechanism doesn't tell you how strong that mechanism is, relative to the other forces that determine the outcome. People who don't realize this get miffed when other people don't accept arguments that seem tautologically true to them---it comes off as willful ignorance, when it's actually that the other person is trying to discuss the sum while you're arguing for the vector.

• It's also tough when people talk about their own experiences. Everyone has lived one causal chain. If I say, 'grad school will ruin your mental health', that is inarguably true for me. But it might also be true that, on average, it makes people's lives better.