The question of proof in psychoanalysis is as old as psychoanalysis itself. The 1895 ‘Project’ aims at being a project of scientific psychology. The Interpretation of Dreams purports to be a science and not a fantastic construction, a ‘fine fairy tale’, to use Krafft-Ebing's remark, hurled at Freud at the close of one public presentation. Each of Freud's didactic works – the Introductory Lectures, the New Introductory Lectures, and the Outline – represents a new effort to communicate to the layman the conviction that psychoanalysis is genuinely related to what is intelligible and what claims to be true. And yet, psychoanalysis has never quite succeeded in stating how its assertions are justified, how its interpretations are authenticated, how its theory is verified. This relative failure of psychoanalysis to be recognised as a science results, I think, from a failure to ask certain preliminary questions to which I devote the first two parts of my essay; the third part is an attempt to reply directly to the original question.

The criteria for ‘facts’ in psychoanalysis

The first question concerns what is relevant as a fact in psychoanalysis. We may begin by noting that traditional discussions about the epistemological status of analytic theory take it for granted that theories consist of propositions whose role is to systematise, explain, and predict phenomena comparable to those which verify or falsify theories in the natural sciences or in human sciences which, like academic psychology, themselves adopt the epistemology of the natural sciences. Even when we are not dealing with a narrow empiricism which does not require a theory to be directly validated by observables, we nevertheless continue to ask the same questions we would put to an observational science. In this way, we ask by what specific procedures psychoanalysis connects this or that theoretical notion to definite and unambiguous facts. However indirect the verification process may be, definitions must become operational, that is, they must be shown to generate procedures for verification and falsification. And this is precisely what is in question: what in psychoanalysis merits being considered as a verifiable fact?

My thesis is that psychoanalytic theory – in a certain sense which will be described in the second part of this essay – is the codification of what takes place in the analytic situation and, more precisely, in the analytic relationship.