Strikes me that arguments over whether structural racism "caused" this or that event mirror arguments over whether climate change "caused" this or that event.

Wrote about that way back in 2012. https://grist.org/climate-change/did-climate-change-cause-the-colorado-wildfires/

Did climate change ’cause’ the Colorado wildfires?

The question of whether climate change "causes" certain events drives David Roberts a little nuts, but he tries to tackle it once and for all.

Grist
The analogy I always use is, imagine you turned gravity up by 5%. More people would stumble. Would any given stumble be "caused" by the increase in gravity? Not proximally: there's always something closer, a pebble or uneven sidewalk, to point to that does more explanatory work.
Nonetheless, an increase in gravity would cause a rise in stumbles. You might call this distal causation, a change in background circumstances linked to a regular & predictable pattern of effects. It's a cause that *raises the likelihood* of an effect.
Most people intuitively understand this difference, but lots of the discourse around both climate change & structural racism involves people pretending you're talking about proximal causation when you're talking about distal causation.
@drvolts Also an apt analogy for COVID sequelae.
@drvolts
I have to think about analogizing for racism, but I use the Tocoma Washington Bridge failure as a visual example of driven oscillation, i.e. what happens when you keep pumping energy into an oscillating system.
Global warming is pumping ever more energy (heat) into the oscillating atmospheric systems of highs & lows & crashing, storm causing boundaries, driving higher highs, lower lows, bigger gradients, more violent storms.
https://youtu.be/XggxeuFDaDU
Tacoma Bridge Collapse: The Wobbliest Bridge in the World? (1940) | British Pathé

YouTube

@drvolts
Plus there's still the problem that for lots of people, racism = personal animosity. That is how mass entertainment has portrayed it for generations.

No animosity? Not racist! Problem solved!

@drvolts and if we know occurrences of event y become a given percentage more likely, then congratulations you've just signed up for that many additional occurrences across your sample data set.

So x more category 5 hurricanes, flood events, whatever were caused though any one occurrence can't be said to have been caused.