The causal uncertainty model revisited