Inferring and Explaining

60 InferrIng and exPlaInIng Ferdinand von Herba was really a contem- porary of Semmelweis. He, just like his young friend and colleague, was very much a disciple of the new pathological and diagnostic tech- niques being taught by Rokitansky and Skoda. He took this background and used it to remark- able advantage by almost single-handedly beginning the specialty of dermatology. Herba’s role in Semmelweis’s story is twofold. He was friend and confdant to Ignác as he sufered through the childbed fever epidemic, and he was the frst to fnally share the discovery with the medical community, in both a professional presentation he gave and a short medical jour- nal publication. What Was Then Known I want to sharewith you a rather lengthy excerpt fromamarvelous book on Semmelweis, Te Doc- tors’ Plague , by the contemporary medical histo- rian Sherwin B. Newland. You will notice that Dr. Newland summarizes Semmelweis’s back- ground knowledge of the magnitude and details of the childbed fever epidemic by introducing each short paragraph with the phrase “obser- vation no.” It is almost as though Newland is consciously schematizing the preliminary evi- dence in terms of our inference-to-the-best- explanation (IBE) recipe, and we can mentally substitute our convention of replacing the “observation no.” with the appropriate e and subscript. Observation no. 1: Te same number of deliveries took place in the hospital’s two obstetrical divisions, usu- ally between 3,000 and 3,500. Te only diference between themwas that deliveries in the First Division were carried out by doctors and medical students and those in the Second Division bymidwives and students of midwifery. In the First Division, an average of 600 to 800 mothers died each year from puerperal fever; in the Second Division, the fgure was usually about 60 deaths, one-tenth as many. Observation no. 2: Although childbed fever raged violently in the First Division, there was no such epidemic outside the hospital walls, in the city of Vienna. Te mortality of home delivery, whether by midwives or private doctors, was low. Even when they gave birth in alleyways and streets, the so- called Gassengeburten , mothers who self-delivered rarely died. Observation no. 3: Despite a general impression to the contrary, the decades of carefully kept statistics at the Allgemeine Krandenhaus showed that neither the incidence nor the mortality of puerperal fever was related to the weather, as epidemics ofen were. Observation no. 4: Greater trauma during deliv- ery appeared to increase the likelihood that a mother would develop puerperal fever. Tis was true of no other epidemic disease. Observation no. 5: Closing the ward for a period of time would always stop the mortality. When mothers were delivered elsewhere during that time, they did not get sick. Observation no, 6: The infant delivered of a woman who subsequently died of puerperal fever would not infrequently die of a fever similar to its mother’s. In such cases, the fndings on autopsy were similar to those identifed in the mother. 3 Diferences in the Divisions Consider the confusion in the two kinds of disease etiology we have discussed. Semmel- weis’s superiors endorsed simultaneously the