I don’t really think this is true. See this historical article:

http://math.bu.edu/people/aki/14.pdf

He was drawn to logic by discussions with Feferman and worked unsuccessfully on other problems in logic before moving to independence questions in set theory.