A number of closed-loop postulations to explain motor skills learning and performance phenomena have appeared recently, but each of these views suffers from either (a) logical problems in explaining the phenomena or (b) predictions that are not supported by the empirical evidence. After these difficulties are discussed, a new theory for discrete motor learning is proposed that seems capable of explaining the existing findings. The theory is based on the notion of the schema and uses a recall memory to produce movement and a recognition memory to evaluate response correctness. Some of the predictions are mentioned, research techniques and paradigms that can be used to test the predictions are listed, and data in support of the theory are presented.

We argue herein that typical training procedures are far from optimal. The goal of training in real-world settings is, or should be, to supporr two aspects of posttraining performance: f a ) the transfer rhat training ro related tasks and altered contexts. The implicit or explicit assumption of those persons responsible for training is that the procedures that enhance performance and speed improvement during training will necessarily achieve these two goals. Howelver, a variety of experiments on motor and verbal learning indicate that this assumption is ofien incorrect. Manipulations that maximize performance during training can be detrimental in rhe long term; conversely, manipulations chat degrade rhe speed of acquisition can support the long-term goals of training. Thefact that there are parallel level Of Performance in the long term and (b) the capability to findings in the motor and verbal domains suggests that princiPies of considerable generali0 can be deduced to upgrade training procedures.Over the past several years, through the normal Process of conducting our own individual research programs (in movement learning and human memory, respectively), and as a consequence of listening to and reading reports of each other's work, we have repeatedly encountered research findings that seem to violate Some basic assumptions about how to optimize learning in real-world settings. For example, increasing the frequency of information presented to learners about performance errors during practice improves performance during training, yet can degrade performance on a test of long-term retention or transfer. Increasing the amount of task variability required during practice, in contrast, depresses performance during training, yet facilitates performance on later tests of the ability to generalize training to altered conditions. Such findings challenge common views of skill learning. Compared with some baseline training condition, how can a factor that enhances performance in practice interfere with retention or transfer performance? Even more intriguing, how can another factor that deGeneral Article grades performance in practice enhance retention performance?These findings-and others we discuss below-are oberal different verbal and motor tasks, and the theoretical motivations guiding those research efforts are often different as well. Taken together, however, these findings tained from diverse research paradigms that employ sevsuggest that certain conce~tualizations about how and when to Practice are at best incomplete, and at worst incorrect. These findings also have some theoretical hplications with respect to the processes involved in practice, particularly as they relate to the acquisition of realworld skills.In this article, we first describe what we regard as Some of the viewpoints, assumptions, and paradigms that, implicitly or explicitly, have provided the foundation for the typical procedures that guide practice and skill acquisition. These views of learning, though flawed in our opinion, have had a strong in...

Theoretical accounts of the speed-accuracy trade-off in rapid movement have usually focused on within-movement error detection and correction, and have consistently ignored the possibility that motor-output variability might be predictably related to movement amplitude and movement time. This article presents a theory of motor-output variability that accounts for the relationship among the movement amplitude, movement time, the mass to be moved, and the resulting movement error. Predictions are derived from physical principles; empirical evidence supporting the principles is presented for three movement paradigms (single-aiming responses, reciprocal movements, and rapid-timing tasks); and the theory and data are discussed in terms of past theoretical accounts and future research directions. Examining the current level of understanding about the production and control of motor responses, many would no doubt be tempted to say that we have not come very far since the early writings of Woodworth (1899) and Hollingworth (1909). These writers were concerned with the basic laws of limb movements (analogous, perhaps to the basic laws of motion that were the cornerstone of physics) that denned the relationship between the simplest aspects of motor

Previous analyses of knowledge of results (KR) and motor learning have generally confounded the transient performance effects as shown when KR is present and the relatively permanent (i.e., learned) effects that we argue should be evaluated on a transfer test without KR. In this review, we classify investigations according to this distinction, and a number of new relations emerge between KR and both learning and performance. In addition to the motivational and associational roles of KR, we emphasize that it also acts as guidance, enhancing performance when it is present but degrading learning if it is given too frequently.

Evidence from nonspeech motor learning suggests that various principles may interact with each other and differentially affect diverse aspects of movements. Whereas few studies have directly examined these principles in speech motor (re)learning, available evidence suggests that these principles hold promise for treatment of motor speech disorders. Further research is necessary to determine which principles apply to speech motor (re)learning in impaired populations.

Relative frequency of knowledge of results (KR) is the proportion of KR presentations to the total number of practice trials. Contrary to predictions from most traditional motor learning perspectives (e.g., Adams, 1971;Schmidt, 1975; Thorndike, 1927), recent evidence suggests that, compared with practice in 100% relative frequency conditions, practice with lower relative frequencies may be beneficial to longer-term retention and learning, but detrimental to practice performance. Three experiments are reported in which the effects of variations in acquisition KR relative frequency were examined. Experiment 1 showed that a markedly reduced KR relative frequency during practice was as effective for learning as measured by various retention tests, compared with a 100% KR practice condition. In Experiments 2 and 3, when the scheduling of KR was manipulated so that the number of KR trials was systematically lowered across practice, a reduced average relative frequency enhanced learning as measured by a delayed no-KR retention test (Experiment 2) and a retention test in which KR was provided (Experiment 3). Results are inconsistent with predictions from an acquisition-test specificity hypothesis and conventional motor learning theories and thus suggest a revision in the principles governing the role of KR for motor learning. Empirical support is provided for the KR guidance hypothesis (Salmoni, Schmidt, & Walter, 1984) and for various encoding-retrieval operations associated with spaced retrieval practice. Possible learning strategies invoked by relative frequency and other related practice variations are discussed with respect to response consistency and the development of intrinsic error detection mechanisms.

No abstract

No abstract