Hanly C
Int J Psychoanal. 1992 Summer;73 ( Pt 2):293-301.
It is argued that clinical observations can be used to test psychoanalytic theories. Psychoanalysis cannot use controlled experiments but it can use the standard canons of inductive reasoning to test the reliability of conflicting theories. Conflicting hypotheses concerning narcissism and object relations are analysed to identify crucial observations that can be used for testing. If these experiments in theory testing are reliable they show that psychoanalysis is not limited to being a hermeneutic discipline but rather has legitimate claims to be a natural science.