rom the early 1950s until the early 1960s, Wallace Russell and I directed a project aimed at examining the role of language in behavior. While that project involved most of the staff of the psychology department at Minnesota at one time or another and concerned itself with everything from clinical protocols to nonsense syllables, the long-term, persisting core of the project turned out to be the studies that Russell and I performed with word associations. Here we examined the effects of established associations on learning, recall, recognition, perception, and problem solving. Everywhere we looked we seemed to find powerful relationships. Then we proceeded to establish associations de novo and repeated these studies with materials constructed entirely in the laboratory so we could control the histories of the associations that we were dealing with. (An account of these studies is given in Jenkins, 1963Jenkins, , 1970 Those were very exciting days for us. When the research program began, there was virtually no experimental evidence for mediation theory, although such a theory was widely applied in the analysis of higher mental processes (e.g., Osgood, 1953). By the time the program was over, media-