1990 is the year the dimensional overlap model (DO) first saw the light of day. Paul Fitts (Fitts and Seeger, 1953; Fitts & Deininger, 1954) had identified a family of tasks in which performance could not be accounted for by either the stimulus properties alone, nor the response properties alone – the results were clearly the outcome of an interaction between the two. He called the phenomenon “stimulus response compatibility effect”.
We proposed that this interaction occurred at two levels: the first, where categorical stimulus and response sets are perceived as corresponding or not. We called the basis of this correspondence the dimensional overlap between the sets. Dimensional overlap was defined as the degree to which attributes in the two sets are perceptually, structurally or conceptually similar. It is, therefore, a property of the mental representation of sets. When dimensions in a stimulus and response set overlap, a second level comes into play: the instructions that map individual stimulus elements onto individual response elements.
This early form of the model already included a taxonomy with four ensembles – which was later expanded to the final eight-task taxonomy in Kornblum (1992), It had a rudimentary architecture, and the outlines of a processing model that combines automatic and control mechanisms.