Would you like that drink up or on ice?

That is, the observed covariance between indicator y₁ and y₃ would mirror the understandable causal behavior of the corresponding underlying latents.

The virtue of the yaller dog is his bark. To hear that bark and not see the barker would make one imagine that he did the biting for all dogdom: but seeing the source one only says, “Poh, it's the yaller dog !”

Let me see your portous, gentle sir John.