Semmelweis ideas were probably rejected because there were no theroretical framework which could explain how disease could be carried around by doctors or nurses who were not themselves infected (bacteria were not yet dicovered). It was not easy for the medical establishment to accept a theory that relied on the existence of a completly unknown and invisible substance.
Also, the post-napoleonic Austria-Hungary was an extremely conservative society. Revolutionary ideas and paradigm-changes were definitiely not the positive buzzwords they are today.
Here is a wonderful quote by the emperor: (http://www.archive.org/stream/austriahungary00legeiala/austr...): Hold to that which is old, for that is good; if our ancestors were pleased with it to be so, why not we? New ideas are to-day being advanced of which I do not nor ever shall approve. Hold them in suspicion and keep to that which is approved. I have no need of learned men; I want faithful subjects. Be such! that is your duty; he who would serve me must do what I command. He who cannot do this, or who comes with new ideas, may leave us; if not, I shall send him.
Not the best environment in which to promote radical new ideas!
In hindsight, Semmelweis opponents made a great error by rejecting a theory which had strong empirical evidence just because it didn't have a theoretical underpinning. However I fail to see the connection to Dijkstras formal methods. Is there really overwhelming empirical evidence for the superiority of formal methods in software design? No, it actually seems Dijkstras argument for formal methods is that it is a "time-honoured tradition", a line of reasoning akin to that of the the emperor and the opponents of sterialization.
http://www.youtube.com/watch?v=vsafOyAdayg&feature=PlayL...
If you've never seen "The Day the Universe Changed," you're in for a treat.