The Tolerance Inversion

Then a dataset of more than 600,000 people walks in and ruins the story.

The analysis is plain about it. People who prioritize free expression are more accepting of marginalized groups, not less. The correlation runs in the exact opposite direction from the narrative — the people most willing to let speech they hate exist are, on average, the people most willing to let people they don't understand exist. Same muscle. The openness doesn't stop at ideas and start discriminating at humans. It generalizes.

Sit with the discomfort, because it isn't in the finding. It's in what the finding says about the one who assumed otherwise.

The comfortable fiction was never really about the bigots. It was about you. Believing that everyone who disagreed with your speech preferences was secretly a bigot let you skip a harder examination: whether your own impulse to restrict was tolerance or its costume. If the other side is arguing in bad faith, you never have to ask whether you're the one who finds certain people easier to silence than to sit with. The accusation was a door you closed so you wouldn't have to walk through it.

Here's the machinery underneath. You ran on an assumption of scarcity — that expression is a fixed quantity, that every platform defended for someone you find dangerous is a platform taken from someone vulnerable. Speech as a zero-sum pie. Under that model, tolerance requires restriction; you protect the margins by policing the center. It feels like care. It's the logic of a clenched fist.

The data describes a different physics. The people who hold the field open hold it open for everyone, including the marginalized, because that's what holding a field open is. You can't run an exception engine that welcomes the speech you like and the people you like while excluding the speech you fear, and expect the welcome to stay intact. The nervous system doesn't compartmentalize that cleanly. Openness is a disposition before it's a policy, and dispositions don't read the room.

This is the part coherenceism would name without flinching: resonance expands a field; dominance contracts it. The assumption of scarcity was the distortion all along. When you decide that protecting the vulnerable means controlling who gets to speak, you've already accepted that the field is small and someone has to guard the gate — and you've quietly nominated yourself. The free-speech prioritizers in that dataset made the opposite bet. They assumed the field was big enough to hold the disagreement, and the tolerance came along for free, because it was the same gesture the whole time.

None of this absolves the actual bad-faith actors. They exist. People do launder cruelty through high principle, and you've correctly clocked some of them. But you used the worst examples to write off the disposition, and the disposition turns out to be the one most correlated with the welcome you claim to want. You profiled an entire stance by its most cynical practitioners — which is, notably, the exact move you accuse them of making about the people they fear.

The invitation isn't to start defending everything anyone says. It's smaller and more annoying than that. The next time someone defends a right you'd rather restrict, notice the speed at which you reach for they must be hiding something. That speed is the tell. It's the nervous system protecting a story it finds more comfortable than the data — the story where intolerance is always the other guy, and the door you closed was for everyone's good.

It wasn't. It rarely is.