Cognitive psychologists argue that over time, task-related knowledge is organized into information-rich representations or chunks, a process termed chunking. Dechunking refers to when well developed chunks are broken down into a number of smaller chunks or bits of information. Chunking is a fundamental building block of human memory and underlies the learning of perceptual and motor skills. Dechunking is a mechanism that potentially underlies suboptimal performance. In this entry, the process by which information is chunked during learning is explained in general terms and in relation to both perceptual and motor skills. The process and implication of dechunking are then briefly discussed.
To operate effectively, humans must utilize information in the performance environment, but the human information-processing system has limitations. Specifically, the capacity to store and retrieve verbally coded, declarative knowledge is finite and small, and serial processing of such information is time consuming. Despite these limitations, people can successfully execute perceptual–motor skills in environments that are often complex, dynamic, and time constrained. Indeed, experts are defined by their ability to overcome the limitations of the information processing system. Their decisions depend upon rapid processing of large quantities of perceptual information. Their motor skills are executed in a smooth and automated fashion. Chunking underlies these capabilities.
Stage models of skill acquisition propose that early in learning bits of information pertaining to a task are stored, retrieved, and processed as declarative knowledge by a short-term information processor (working memory), which interfaces with long-term memory. With repeated on-task experience, meaningful associations emerge between recurring bits of information that are grouped together and coded as chunks. Over the course of practice, between-chunk associations develop and higher and higher level chunks are compiled, until large amounts of the information pertaining to a task are hierarchically organized into a single representative chunk. The demand on information processing resources lessens as each level of the hierarchical structure is formed, until higher level chunks take a procedural form that no longer requires the retrieval and processing of declarative knowledge. In the domain of sport and exercise, chunking takes two distinct forms: the chunking of perceptual information for response selection, and the chunking of the motor information needed to organize and execute a successful response.
The strategies and rules inherent within dynamic activities like team sports generally result in the availability of structured perceptual information, such as the relative position of teammates and opponents or the mechanics of an opponent’s movements. Visual elements that meaningfully correlate gradually become recognized as distinct patterns. With increased exposure, relations between these patterns emerge and so they are compiled together, and from then on recognized as a single pattern. Empirically, the chunking process has been demonstrated by findings that show experts have superior capacity to recall patterns of game play after only a brief presentation of a structured, rather than transitory, unstructured game scenario. Chunking allows experts to quickly and accurately recognize patterns from within the unfolding events and this, in turn, permits both the rapid and accurate selection of an appropriate motor response.
A motor response is the outcome of the ordered execution of a sequence of movement components. In the early stage of motor learning, the sequencing and control of independent movement components is typically controlled at a conscious level, placing high demands on the available information processing resources. Consistent repetition of a pattern of movement through practice causes adjacent movement components to coalesce into chunks. Chunks in motor learning are defined by pauses between successive movement components. For example, sequential key pressing experiments show that latencies between successive key presses decrease with practice in a clear and consistent fashion such that longer pauses demarcate chunk boundaries. Pauses between adjacent motor chunks result in a segmented or jerky motor action. Continued practice builds a hierarchical structure, in which fewer chunks comprising longer sequences of movement components exist. As a result, the movement is both less jerky and places fewer demands on information processing resources. Eventually, all movement components are incorporated into a single representative motor chunk, which, when selected, results in the smooth, automatic production of complex sequences of movement.
Compared to the concept of chunking, dechunking is a less well established phenomenon. As chunking is associated with the learning or progression of skills, dechunking may be used to explain occurrences of skill regression or skill failure. Declarative knowledge accrued during learning likely remains retrievable from long-term memory after it has been chunked at a higher level. The retrieval and processing of previously chunked declarative knowledge represents the dechunking of optimal and automated representations and regression to a less efficient and less fluid mode of control. In the competitive environment, contingencies such as performance anxiety may trigger dechunking and result in skill failure. Alternatively, in the training environment athletes may purposefully attempt to revise the composition of high-level representations by first dechunking and then rechunking information.
In summary, the chunking of information is a consequence of the practice that underlies skill learning. It allows well-practiced performers to operate proficiently despite the temporal and capacity constraints of the information processing system. Dechunking describes a return to the retrieval and processing of hitherto chunked declarative knowledge.
- Anderson, J. R. (1982). Acquisition of cognitive skill. Psychological Review, 89, 369–406.
- Gobet, F., & Simon, H. A. (1998). Expert chess memory: Revisiting the chunking hypothesis. Memory, 6, 225–255.
- Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81–97.
- Rosenbaum, D. A., Kenny, S. B., & Derr, M. A. (1983). Hierarchical control of rapid movement sequences. Journal of Experimental Psychology: Human Perception and Performance, 9, 86–102.