Most clinical problems can be boiled down to the following practical syllogism: "If a patient has a condition p, then he should be treated with q, r, or t or whatever combination of them. The patient X has the condition p. Therefore, the patient X should be treated with q, or r, or t, or whatever combination of them". It is evident that the conclusion of this syllogism is a result of two different kinds of knowledge: first, medical knowledge understood as general and universal knowledge of health, disease, and treatment which is contained in standard medical textbooks and which, according to the present fashion, is called evidence based medicine; and second, clinical knowledge which is specific knowledge of a particular patient in terms of his unique narrative identity. (Recently, this kind of knowledge has been named narrative based medicine.) Acquisition and application of medical and clinical knowledge are governed by ethical rules. The basic rules of human subject research ethics are presented and discussed. I conclude that even a perfect evidence based medical knowledge is not enough to make a sound clinical judgement. Because every human individual may have a specific reaction to a drug, each treatment is always experimentation on the human being and basic rules of human experimentation must be applied. One of the fundamental rules of clinical experiment is respect for moral autonomy of the patient which is reflected in the moral, legal, and clinical doctrine of informed consent; another, the principle of an acceptable risk-benefit balance. Both principles are examined in the context of psychiatric treatment.