A physician investigating a patient’s complaint will try to form an image of the problem by conversing with him. This is the anamnesis. Next he forms a hypothesis, the possible diagnosis. After that he will attempt to verify or rule out his hypothesis by further examination. This process of verifying or ruling out a diagnosis is called validation. The word validation will be used here in this meaning. In addition, the concepts of reliability and predictive validity will be used, as does Kendell, as statistical-scientific concepts. The reliability with which for example a certain diagnosis can be determined is the measure in which one can be sure that that diagnosis is indeed correct. Predictive validity is the measure in which the determination of the diagnosis allows prediction of future events such as for instance the determination of a prognosis and how it may be influenced by treatment. Predictive validity is crucial in diagnostics. Its accuracy depends on the reliability of the diagnosis. So high reliability of the diagnosis is a necessary condition for a good predictive validity. But predictive validity is not determined by the diagnosis alone, so in itself not a sufficient condition for it.

One of Szasz’s arguments in support of his view that mental illness does not exist is the problem of validation. When a physician suspects an organic aberration he has all sorts of physicochemical methods at his disposal by which to demonstrate this aberration objectively. In contrast, the psychiatrist has only his subjective judgment to pose along with or opposite the patient’s. No objective criterion for proof is possible. When the psychiatrist’s and the patient’s opinions differ, the psychiatrist’s is decisive, not because he is right – that cannot be proved – but because power is on his side. So here there is no scientific examination of the nature of things, but a – moral – confirmation of power of the one over the other, which has no relation to disease, according to Szasz.

In order to judge the value of this argument it is necessary to examine it more closely and compare the validation process in somatic medicine and psychiatry. One immediately noticeable difference is that physicochemical methods of validation are not applicable when, as in the majority of psychiatric disorders, no organic aberrations are known. Advocates of a materialistic biomedical disease concept who value only physicochemical findings are justified in positing that validation regarding most psychiatric disorders is not possible. However, those who do not maintain such an absolute contrast between physical scientific insight on the one hand and every other insight on the other can ask how validation in psychiatry works, and compare this with validation in somatic medicine.

Below I will first examine the process of validation in somatic medicine more closely. Afterwards I will do the same for psychiatry. Finally I will compare the two.