ABSTRACT
A long-standing, unsolved puzzle in motor skill acquisition is the problem of transfer: How does experience in a familiar motor context facilitate performance in a novel situation? Since Thorndike’s (1906; Thorndike & Woodworth, 1901) classic work at the turn of the century, researchers have assumed that the solution to the puzzle can be found in a search for identical elements. More specifically, the traditional assumption is that transfer depends on the extent to which elements of the training context-the nature of the environmental stimulus, the patterns characterizing the motor response, or the mental representations that support motor knowledge-are similar to elements of the performance context (e.g., Anderson & Singley, 1993). Concepts like stimulus generalization, motor equivalence, and learning via analogous mental elements have each had their heyday (Adams, 1987). However, after nearly 100 years of study, researchers still lack a satisfactory theory of transfer in motor skill acquisition. One of the impediments to progress is that the identical elements approach and the experiments that stem from it cannot, in principle, explain adaptive, functional motor skills. The notion of identical elements in simple associative learning is far too static and narrow to account for the step-to-step, momentto-moment variability that characterizes real, everyday motor skills.
