Two different kinds of uncertainty occur in the clinical process. In the diagnostic phase questions related to categories must be answered, and information concerning symptoms and etiology are evaluated and compared with patterns known from textbooks and general experience. Accordingly, the uncertainty is related to observation, definitions of categories, and usage, and it may be demonstrated by kappa coefficients or similar measures of diagnostic reliability. In the treatment phase, on the other hand, uncertainties are related to probabilities - that is, the quantitative relationships between the categories disorder/treatment/effect. This kind of uncertainty may be described by p values and confidence intervals. The kinship between these uncertainties is analyzed, and improvements in nosography are suggested, to reduce their influence.