[I]t may be that we have to admit that causal influences do go faster than light. … An ‘aether” would be the cheapest solution. But the unobservability of this aether would be disturbing.