Concepts precede facts, and the hypotheses of psychoanalysis are foreign to the way most people think: it is not known, Freud says, whether they should be regarded as 'postulates or as products of our researches'. To guess (which translates the German verb: 'zu erraten') may be the only way to represent to oneself what is happening, because what is happening happens in the transference; and transference is the only formation of the unconscious the patient can tell us nothing about. To guess and to construct are synonymous and, in our constructions, we communicate our guessing to the patient using a kind of intermediate language, which is the uncanny language of transference neurosis and, at the same time, the language of conceptualisation; that language acknowledges the strangeness and foreign nature of transference. When it does not, theory then becomes coded and lifeless. And its concepts are in danger of resembling what the Editorial of the first issue of the 'International Journal' in 1920 called 'catchwords'. Theory should remain for us, Freud said, 'a stranger one has not invited'.