ABSTRACT
The only way of characterizing diseases so that each disease will have a universal necessary cause is by defining them in terms of their causes. For example, the tubercle bacillus is universally necessary for tuberculosis only because ‘tuberculosis’ is defined as infestation by this bacillus. One cannot be certain that no one (perhaps Fracastoro, Paracelsus, Ugo Benzi of Siena, or Sinan ibn Thabit ibn Qurra) employed such characterizations before 1850 although, to date – loose talk about disease specificity notwithstanding – no textual evidence has been presented that anyone did. One reads that Fracastoro envisioned defining diseases in terms of what he called ‘seeds’. But unambiguous evidence has yet to be provided and, even if this had been his intent, in the absence of any empirical means of distinguishing among seeds (or, indeed, even of demonstrating their existence) any such attempt would have been exactly as useless as defining diseases in terms of assorted evil spirits; vacuous definitions explain precisely nothing. However, this much is absolutely certain: Ignaz Semmelweis was the first person to have adopted such definitions in publications that are linked by citations directly into the etiological research programme. The only way to dispute this fact would be to exhibit citations in core works of the programme to writers before Semmelweis who used this approach, and there simply are none.1
