Well, I guess it probably stems from something like this. We knew about gel electrophorsis, to sort proteins by size. We knew that we can use antibodies to detect proteins at the cell surface and in immunoblotting. So one could envision that we could use the antibodies to detect the protein in the gel. hence western blots. A PhD student probably then spent 3 years playing around with the idea until he got something that worked.
So in effect it was an idea stemming from ideas that were already in place. well, I guess that is the order in which it happended, I dont know, though once you understand those two concepts, putting them together is not so hard. So it was a kind of evolution, just one that was slightly bigger than others.