From his first attempts to explain attachment phenomena in the 1940s through his "Attachment and Loss" trilogy (Bowlby, 1969/1982, 1973, 1980), John Bowlby reformulated the theoretical underpinnings of attachment theory several times. He initially attempted to explain attachment phenomena in psychoanalytic terms. Then he invoked ethological theory in the explanation of how and why people behave as they do in close personal relationships. The mature theoretical framework that he presented between 1969 and 1982 in the attachment and loss trilogy retained strengths and insights, ultimately situating them within an overarching control systems framework. This article describes key stages in Bowlby's theoretical development, with particular emphasis placed on the emergence of control systems theory as a cornerstone of the mature theory. It also compares Bowlby's control systems approach to contemporary cognitive science approaches. It concludes by suggesting how Bowlby's control systems formulation could evolve along the path opened up by contemporary work in computational modeling and how it could benefit by doing so.