Medical decisions, including diagnosis, prognosis, and disease classification, must often be made on the basis of incomplete or unsatisfactory information. Data which are essential to the care of one patient may be unobtainable for technical or ethical reasons in another patient. For this reason the principles of controlled experimentation may be impossible to satisfy in human studies. In this paper, some formal aspects of medical decision making are discussed. Special operators for the intuitive concepts of certainty, demand, and effort, akin (...) to the operators of modal logic, are used to accommodate the technical and ethical limitations on human studies. Theorems are stated and proved which show how this system handles incomplete information. The embryogenesis of the human heart is presented as a sample problem in classification. (shrink)