In the standard serial reaction time (SRT) experiment, subjects are required to respond rapidly to a structured sequence of visual targets. Evidence that subjects have acquired knowledge of the structure is obtained by modifying the structured nature of the sequence and noting whether reaction times increase. In the dual-task SRT experiment, a "secondary" tone-counting task is introduced, and the extent to which learning of the "primary" target sequence is compromised is noted. Here we present data that strongly imply that while the psychologists who designed this "dual-task" experiment may have viewed it this way, this may not be the best way to characterize it. The suggestion is that this "duality" is illusory and that we should probably be treating the tone-counting task as a potential source of additional patterns of covariation in a complex, multicomponent display and not as a "secondary," attention-diverting factor.