You do not currently have permission to view the full text for this document. Please sign in to an account with an active subscription or contact support.
According to methodologists,2 a genuine is distinguished from a pseudo-science by the fact that it is possible to say, not only what evidence would prove its theories true, but also what evidence would prove them false.
Our critics argue that, like astrology, psychoanalysis fails to pass this test because we appear to them to have a limitless capacity to defend ourselves against negative evidence by means of additional hypotheses. There can be no disproof, they say, if a patient's denial of an interpretation can be taken only as evidence of a 'resistance'; if his assertion that the opposite of what is said of him is true can be explained away in terms of 'ambivalence' and the coexistence of conscious and unconscious impulses; and if the analyst, when himself convinced that he should have made a different interpretation, fails to withdraw the first one on the ground that, owing to 'over-determination', it is still right at a deeper level.3
Now since the technique of proof and disproof is most advanced in the physical sciences, it might seem natural to take them as our model in seeking to confound our critics. We might try, for example, to show that our theory of the unconscious, like the physicist's model of atoms, enables us to predict observable behaviour, the non-appearance of which would disprove the theory, that is, expose it as useless.4 But however successful such experiments might be, they would not I think provide an adequate defence of what is most vital to us: namely, that specialized form of anthropomorphic reasoning on which, I believe, the whole structure of analysis ultimately rests. This no longer has any parallel in physics—except in so far as models of atoms still contain faint traces of animistic projections of a tactual kind. To prove that, in psycho-analysis, anthropomorphic reasoning can still lead to reliable results is not, I think, sufficient. We have to examine the inherent conditions under which it is reliable or not.
The essence of all