Hierarchical versus nonhierarchical models of movement sequence control: a reply to Klein

Klein (1983) has proposed a nonhierarchical logogen model to account for response timing and error data from a rapid finger-tapping task reported by Rosenbaum, Kenny, and Derr (1983). On grounds of parsimony, Klein's model cannot be easily distinguished from the hierarchical model proposed by Rosenbaum et al. However, the two models can be distinguished on other grounds. First, the logogen model cannot account for results from a second experiment reported in the Rosenbaum et al. article that the hierarchical model can account for. Second, the logogen model gives rise to a preponderence of incorrect response orders, whereas the hierarchical model does not. Third, the logogen model, unlike the hierarchical model, cannot account for the fact that people can produce complex, extended sequences of responses from memory.