the behavioral sciences have commonly insisted upon certain arbi-trary methodological restrictions that make it virtually impossible for scientificknowledge of a nontrivial character to be attained.