Merlin C. Wittrock was a friend and colleague who influenced me and many other contemporary mathematics educators on how students learn. In this article I summarize my interactions with Merl beginning in 1965, and how over the following half-century he influenced my thinking on student learning for understanding and, in turn, on how to design instruction based on generative principles. I situate our initial discussions in terms of behaviorism, mathematical abilities, and mathematical problem solving. Based on Merl's ideas since first expressed in "A Generative Model of Mathematics Learning" in 1974, today it is clear that instruction should involve the stimulation of a student's store of relevant background experiences in relation to information to be learned, so they can construct meaning from it. Only by doing so can the use of such organized structures of previously acquired knowledge and experiences facilitate the understanding of new information. (Contains 1 footnote.)