ABSTRACT
The direct lineage of statistical learning theory began in 1950 with the publication in Psychological Review of Estes' article “Toward a statistical theory of learning.” Before saying more about that I recall, however, that there were a number of earlier attempts to develop a quantitative theory of learning which can easily be cited, but I hasten to say that I am not attempting anything like a serious history of the period prior to 1950. I have used the following highly selective but useful procedure of listing the earlier book-length works referred to in the spate of important papers published between 1950 and 1953, which I discuss in a moment. The earlier references referred to (not counting a large number of articles) are Skinner's Behavior of Organisms, published in 1938, Hilgard and Marquis' Conditioning and Learning , published in 1940, Hull's Principles of Behavior, published in 1943, and Wiener's Cybernetics, published in 1948. These and other earlier works not mentioned played an important role in developing the theoretical concepts of learning used in statistical learning theory and related theories developed by others shortly thereafter. The basic concepts of particular importance are those of association and reinforcement expanded into the concepts of stimulus, response, and reinforcement, and the processes of stimulus sampling and stimulus conditioning. Other important psychologists who contributed to this development and who were not mentioned on the basis of the principle of selection I used are Guthrie, Thorndike, Thurstone, Tolman, and Watson.
