As implied by the tittle of my paper, I think that I have discovered a technical improvement that fits better into the roots of psychoanalysis than into any one of its many current branches. My first point is that the ?abreactive phase? of psychoanalysis was prematurely closed for lack of appropriate tools to investigate the stream of consciousness or, we should rather say, the stream of unconsciousness. After abandoning hypnosis and experimenting with other suggestion techniques (such as the laying of hands), Freud finally settled for "free associations" or the reporting of spontaneous mental contents during the waking state. My second point, based on both historical and personal data, is that free association is a rather clumsy technique, and that it could, and should, be greatly improved. Instructing the average patient, as Freud did, to ?say everything that comes to mind, as it comes,? is something like asking an anxious person to keep calm. In both cases, the instructions are appropriate, but quite difficult to comply with. Stating the goal is not enough; we also need instrumental instructions, that is, we have to teach the patient how to achieve those goals. This leads us to my third point: we have to teach technical procedures for attention management that are adequate for analytic work, rather than relying on the analysand´s intuition to stumble upon some way to free associate.