关键词: 组块, 工作记忆, 模型结构, 容量

Abstract:

Working memory is an important system to store and deal with a mass of information in our daily life. On account of the limitation of working memory capacity, several memory strategies are usually adopted to promote the capacity to satisfy the needs of overloaded information processing. Chunking is one of the most common and effective memory strategies, which has an important impact on daily life and learning. It is a matter of great concern about working memory in psychology. Although previous researches have confirmed that chunking can increase the capacity of working memory, there are debates on the concept of chunking, the relationship between the structure of chunking and development of age, chunking capacity, and its works mechanism. In this case, it is still worth studying about these questions.
Firstly, some previous researches proposed that the strategy of information restructuring is crucial to chunking; the others, however, agreed that the influence of common features among stimuli was more important than strategy. According to this debate of the understanding of chunking processing, the definition of chunking continues to evolve. At the same time, some researchers started to analyze the relationship between the structure of chunking and development of age. In details, there are three kinds of hierarchical organizations in this structure, which involved (1) the individual chunk formed by single and isolated item, (2) the grouping chunk based on the internal or external features' connections, and (3) the superchunk recoded or reconstructed from prior semantic knowledge and experience. Most infants before 14-month-old are in the stage of individual chunk due to their poor prior experience and immaturity of language ability. Until 14-22-month-olds, infants can use higher-order chunking to group items as their language development.
Secondly, another research orientation paid more attention to the theories of chunking and its capacity. Chunking theory proposed a complete system which is responsible to retrieve and encode chunks, but it does not make distinction on whether the chunking works in short-term memory or working memory. This theory also cannot explain how an expert could storage, recall and process amount of information so rapidly. These shortcomings were compensated by template theory. The template theory, however, cannot explain how to group information under the condition of attentional resource limitation due to the lack of attentional system. Besides, embedded-processes model of working memory indicated the processing mechanism of chunking in working memory, which is embodied in the integration and extraction of attention focus in long-term memory. However, some researchers oppose to regard working memory as a part of long-term memory.
In conclusion, the remarkable difference among these theories is the usage of strategies during chunking and whether it could influence on the number of chunks. According to previous researches, we found that the number of chunks in working memory is basically constant. Rehearsal strategy has influence on chunking, which leads to the dynamic change of overall chunks connection, and more flexible to respond to the memory needs of specific information in different situations.
Finally, there are some issues worth concerning in the future, like (1) the role of long-term memory during chunking in working memory, (2) the influence of long-term memory on encoding and storage of chunks in working memory, (3) the features of chunks at the stages of childhood, adolescent and elderly, and (4) to abide the principle of "less is more", the way to improve the efficiency of working memory is to explore a capacity range that can maximize the function of chunking on the basis of the chunk quality, and not to expand chunk capacity in working memory.

Key words: chunking, working memory, structure, capacity