Assuming that A common definition for "orthogonal" is "statistically independent" does it oppose that The meaning of "X is orthogonal to Y" shouldn't imply X can’t have a causal relation with Y?