Evidence that the type 1/type 2 dichotomy in alcoholism must be re-examined.