I don't understand your third point at all. The closest I can come is, if we think of "paradigm" in the broad sense ("disciplinary matrix," meaning a constellation of beliefs, values, formulas, techniques, analogies, exemplars, etc.) as analogous to a body, then a paradigm shift is like changing your body, and in transmitting its information to a future body - its offspring - it can't transmit acquired information? If so, your point seems wrong, so it's probably not what you mean. But if it is what you mean, (a) some information isn't transmitted at all (as with the characteristics in biological evolution that don't get transmitted to the offspring), given that that's what makes paradigm shifts "noncumulative," but (b) where characteristics are transmitted, they most certainly can be acquired characteristics (unlike in biological transmission, where changes come from combining characteristics of two parents and from mutations and biochemical changes). If you're thinking of a paradigm ("disciplinary matrix") being like a species rather than a body, the evolutionary idea (and there's disagreement about this) is that the evolution of species is nonetheless carried on by individuals, who generally don't pass along their genetic information to other individuals who are not their offspring (well, this isn't totally true, but let's say it is) and don't pass along characteristics that they learn or acquire in their lifetime, since these aren't genetically encoded.

Anyway, if your thought is that "preadaptation" has implications for whether or not characteristics can be acquired, I don't follow you. Whether or not a characteristic is acquired rather than transmitted genetically has nothing to do one way or another with the fact that it can be employed in ways that are different from what made it confer an evolutionary advantage on the individual who possessed it, and these new ways may also provide evolutionary advantages.