ABSTRACT
In a previous article we reported a series of studies of the timing of rapid prespecified action sequences in speech and typewriting (Sternberg, Monsell, Knoll, & Wright, 1978; reprinted as Chap. 15 in the present volume). Under pressure to finish quickly, subjects responded to a signal by typing lists of letters or speaking lists of words that had been specified before the signal. Increments in sequence length increased the time to initiate the first response element (latency effect) and also increased the average time from one element to the next (duration effect). Our experi ments permitted us to reject several interesting explanations of the la tency effect, and the generality, robustness, and quantitative simplicity of the two effects encouraged us to develop a model for the latency and duration of rapid movement sequences that explained both phenomena in terms of retrieval of “ subprograms” from a “motor program” for the whole sequence that had been compiled in advance. Our results also permitted us tentatively to identify a response unit in speech-the stress group-in terms of which the observed effects can be most simply de scribed.
