[js8]

I agree, pure empiricism can lead to superstition. If you only learn from experience, and do not have any theory that ensures the consistency of the model, it's easy to infer wrong causal connections.