The article indicates the difficulties with different versions of memory a patient brings in subsequent analytic sessions. The imaginative memory of the patient depends on unifactorial person schemata, on leading through the actual state of emotion, and on the ascription of internal causes. The auditive memory of the analyst depends on multifactorial person schemata, on leading through dynamic and structural concepts, and on the ascription of internal causes too. An approximation to overwhelm the discrepancies lies in translating the patient private personality theory into therapist science language, and retranslating it in patient every day language. Some precautions are discussed.