This paper reviews the last 50 years of experimental and theoretical research on Pavlovian conditioning in animals. It is the history of the movement from simple “spit-and-twitch” psychology to information processing views of associative learning. In 1962, Egger and Miller reported a pivotal study suggesting that information value is important in establishing an effective conditioned stimulus. In the late 1960s, Wagner, Rescorla, and Kamin published historic research papers demonstrating the importance of information value (i.e., predictability of the forthcoming significant event) by showing new phenomena in Pavlovian conditioning: relative cue validity, contingency effect, and associative blocking. The Rescorla–Wagner model came on stage in 1972 to explain these phenomena and successfully predicted new phenomena, although this model had some shortcomings. Subsequent theories of Pavlovian conditioning have challenged to deal with these shortcomings, and the recent theoretical development is linked to computational modeling in a variety of ways.