Caroline Palmer - US grants

The proposed research tests the assumption that human performance in domains such as speech, music, and typing involves the construction of mental representations. A cognitive theory of human performance must combine a) mental structures and units of knowledge with b) cognitive retrieval processes for their organization and implementation in performance. Theories of performance can be tested through study of production errors: performance breakdowns resulting in unintended output. A successful theory will specify the kinds of errors likely to occur, constraints on the forms that errors take, and the conditions that precipitate them. Musical behavior provides a rich domain in which to study cognitive processes underlying performance. With specialized equipment, methodology, and computer software, we can now precisely measure, analyze,and reproduce music performance. Experiments are described that manipulate the type and amount of information available prior to performance, affecting the mental structures and processes used to plan performance. The distribution and context of production errors provides evidence to test: a) the existence of mental plans prior to performance, b) the mental structures and processes from which performance is organized, and c) interdependencies among mental structures and processes. Additional experiments address changes in cognitive processes accompanying acquisition of skilled performance, and whether or not knowledge of performance plans affects music perception. The mental structures and processes underlying music performance may be based on cognitive principles governing other human performance systems as well.

DESCRIPTION (Applicant's abstract): This proposal outlines a framework for skilled performance that focuses on both relative timing and serial order constraints. Most complex actions produced by humans, such as speech and music, have systematic, information-bearing temporal patterns that are not well-understood. Study of production errors as well as event timing in sequence production has documented constraints on both serial order and temporal characteristics of performance, but the findings are rarely considered simultaneously in theoretical accounts. Experiments are proposed that examine relationships among relative timing and serial order in music performance, a complex production domain which offers considerable flexibility in sequence structure and temporal task demands. With specialized equipment and methodology, we can now precisely measure, analyze, and reproduce event timing and production errors in music performance under natural conditions. Another goal is to determine at what point the cognitive representations for action become independent of the effectors on which they operate. What accounts for differences in motor teaming between effectors (hand I finger movements) and across tasks? A final goal is to address learning: the representational structures and processes that allow us to use past experience to improve future behavior. How do memory structures and processes combine with temporal Constraints to predict a shift from novice to expert sequence production in item selection, range of planning, and effector independence? Experiments address these issues in studies of long-term and short-term practice effects, using a combination of naturalistic and experimental paradigms. Outcomes would have significant impact in their potential for instructional methods and their relevance to other sequence production domains.