OBJECTIVE To identify problem solving strategies in general practice. BASIC PROCEDURES Three styles of scientific reasoning were defined and modelled on the medical environment. These models were tested in a simulated doctor-patient encounter. MAIN FINDINGS According to the definitions contained in the models a deductive or hypotheticodeductive strategy could not be discovered. All participants used exclusively the inductive method and, more specifically, the speculative variant. This variant may be best described as a process of iterative pattern recognition. CONCLUSIONS The manifest utilization of the inductive method carries many implications among which are the irretraceability and irreproducibility of the process. The speculative form approaches the conception of art rather than science. This is exactly what most doctors try to tell: medicine is an art. The strategy allows for flexibility and quick response to the patient's needs.