> We have something better for evaluating PLs.

Do tell! How do you evaluate empirical claims (about programming languages) without empirical evidence?