If you’re going to try to formalize the arguments, people aren’t arguing over whether they’re independent variables (or correlated as you say), it’s what they’r

If you’re going to try to formalize the arguments, people aren’t arguing over whether they’re independent variables (or correlated as you say), it’s what they’re conditionally independent based on. If the causal graph is (race —> parents income —> kids education —> kids intelligence), then no shit they’re correlated. But if, conditional on parents income (as a random example, not intended to be exhaustive), race and kids intelligence are independent, then the fact that they’re not (unconditionally) independent is fucked up and needs equity. There’s plenty of evidence of mediating variables like that that are making things inequitable.