What does irreducible complexity have to do with Darwinian evolution? Evolution by mutation and natural selection must proceed by one slight, functional improvement at a time. So how can it build an irreducibly complex propeller motor one step at a time if the motor can't propel at all until all of its parts are in place? It can't. Something else built it.
Behe's argument doesn't assume that none of the other parts could ever be used for anything else. The spring on a mousetrap could be taken and used in some other device. The base with cheese on it could feed a mouse. Several but not all of the parts of a bacterial flagellum--while completely useless as a rotary propulsion machine--are enough for a machine that transports proteins across a membrane. Does this island of functionality provides a credible Darwinian pathway?
Imagine if I told you I could climb to Mars because there exists a natural ladder stretching from here to the red planet. You're naturally skeptical, noting that nobody has ever discovered such a ladder. I then scream, "That's an argument from ignorance! Scientists are finding all sorts of new things all the time! Look! The moon! It's one step along the way! See, everything's falling into place!" The Darwinists' efforts to deny the significance of the bacterial flagellum is strangely akin to this sort of reasoning.