Superior Memory of Experts and
Long-Term Working Memory (LTWM)
An updated and extracted version of Ericsson (in press)
An expert in a given domain of activity, such as medicine, chess, music or golf, is "one, who has acquired special skill in or knowledge about a particular subject through professional training and practical experience" (Webster's, 1976, p. 800). Experts will therefore, by definition, have a greater body of knowledge about their domain of expertise than other individuals. More remarkable is the experts' accurate memory for new experiences in their domain. An elite athlete can, after a sports event, discuss the play-by-play action. Expert chess players can readily recall details of chess positions from their matches in recent tournaments. Early in the twentieth century it was believed that experts were innately talented with a superior ability to store information in memory. Numerous anecdotes were collected as evidence of an unusual ability to store presented information rapidly. For example, Mozart was supposed to be able to reproduce a presented piece of music after hearing it a single time. However, more recent research has rejected the hypothesis of a generally superior memory in experts and has demonstrated that experts' superior memory is limited to their domains of expertise and can be viewed as the result of acquired skills and knowledge relevant to each specific domain.
The most influential research of experts' memories focussed initially on chess experts. In their pioneering studies Chase and Simon (1973) showed superior memory for chess positions by chess experts. Chess players ranging from beginners to international masters were shown a position from an actual chess game (such as the one illustrated in panel A of Figure 1) for a brief time (normally 5 seconds) and then asked to recall the location of all the chess pieces. The ability to recall increased as a function of chess skill. Beginners at chess were able to recall the correct location of about four pieces, whereas international-level players recalled virtually all of the more than twenty pieces.
To rule out that the superior memory of chess experts reflects a general superior ability to store any kind of visual information, Chase and Simon (1973) had chess players recall chessboards with randomly placed pieces (as illustrated in Panel B of Figure 1). With briefly presented random chessboards, players at all levels of skill had the same poor recall performance and were able to recall the correct location of only about four pieces-a performance comparable with that of chess beginners for actual positions from chess games. Further, Chase and Simon (1973) showed that when an actual chess position was shown using an unfamiliar notation (see Panel C in Figure 1), the chess expert was able to display a similar level of superior memory performance after a brief period of adjustment. This result implies that the superior memory of experts is not photographic and requires arrangements of chess pieces that can be encoded using associations to the experts' extensive knowledge of chess. Since Chase and Simon's classic study, investigators have shown that level of expertise is related to superior memory performance for representative stimuli in the associated domain, such as computer programming,, basketball, and dance, and that this superiority is mediated by increased knowledge and domain-specific skills (for a recent review, see Ericsson, Patel & Kintsch, 2000).
Unless one has the knowledge of the expert, it is difficult-indeed, impossible-to grasp the meaningful relations between chess pieces perceived by the expert in panels A and C of Figure 1. If, on the other hand, the availability of knowledge providing meaning to a stimulus is critical to superior memory, it should be possible to demonstrate the same effect in a domain where all adults are proficient, such as language. Human adults are able to recall verbatim meaningful sentences of twenty or more words after a brief presentation (Chase and Ericsson, 1982). An example of such a sentence would be THE WOMAN IN FRONT OF HIM WAS EATING PEANUTS THAT SMELLED SO GOOD THAT HE COULD BARELY CONTAIN HIS HUNGER. If the words of the sentence are randomly rearranged (analogous to Chase and Simon's procedure for generating random chessboards), accurate verbatim recall drops to around six words. An example of a random rearrangement of the above sentence would be WAS SMELLED FRONT THAT HIS THE PEANUTS HE GOOD HUNGER EATING BARELY WOMAN OF SO IN COULD THAT HIM CONTAIN. For random lists of words the recall of subjects is limited by the small number of words they can keep rehearsing (see Repetition and Memory; Working Memory in Humans), and once they stop rehearsal, the words are quickly forgotten. In contrast, once meaningful sentences are understood, their meaning is well retained in long-term memory. For example, during normal comprehension of a text the essential information in each sentence is efficiently stored in memory so it can be integrated with related information presented later in the text (Ericsson & Kintsch, 1995).
Stimuli from an unfamiliar domain of expertise, such as diagrams of chess positions and medical terms, are about as meaningless to most adults as random lists of words and digits. Recent studies have shown that memory for meaningless information can be dramatically improved through training by actively seeking out meaningful associations for the meaningless material. For example, the sequence 671945 can be remembered as 67 being the retirement age and 1945 being the year of the end of World War II (see Mnemonic Devices). Through extended training individuals can acquire memory skills allowing them to increase their memory of briefly presented lists of numbers from an initial level of seven digits to over 80 random digits. Hence, it is possible for regular college students to attain exceptional memory performance after 50 to 200 hours of practice. Laboratory studies of individuals with exceptional memory performance for numbers, names, and pictures reveal that they rely on acquired memory skills, often involving some kind of mnemonics (Ericsson & Lehmann, 1996; Wilding & Valentine, 1997).
The primary goal for all experts is to excel at the representative tasks in their domains. For example, chess experts need to find the best moves to win chess matches and medical experts have to diagnose sick patients in order to give them the best treatment. Unlike the memory experts who attempt to improve their memory performance by acquiring mnemonic techniques through extended practice, chess experts and medical doctors do not deliberately train their memory. Their superior memory ability must thus be an indirect consequence of their improved performance on representative tasks (Vicente & Wang, 1988). Furthermore, experts appear to store task-relevant information in memory when they normally perform representative tasks in their domain, because if they are unexpectedly asked to recall information about a performed task, their memory is typically much superior to that of less skilled individuals. In fact, experts' incidental memory of the relevant information is frequently so good that instructing them to intentionally memorize the information does not reliably improve their memory. For example, when chess experts analyze a position to find the best move, their memory of the position is just as good whether they were informed about an upcoming memory test or not. As part of performing the representative task of selecting the best move the experts encode the important features of the presented information and store them in accessible form in memory. In contrast, when subjects, after training based on mnemonics and knowledge unrelated to chess, attain a recall performance comparable with that of the chess experts, they still lack the ability to extract the information important for selecting the best move. Hence, the remarkable characteristic of expert memory is not just the amount recalled, which can often be matched by training, but the rapid extraction and storage of important patterns and relevant information that allows the experts to perform better the representative task, such as selecting moves in chess (Ericsson et al., 2000).
An analysis of expert performance shows that it is not sufficient to have merely stored the knowledge in memory; it is also critical that relevant knowledge is well organized and can be efficiently retrieved when it is relevant to the ongoing processes. In fact, the principal challenge of expertise is to acquire and organize the vast body of domain knowledge (Chi et al., 1981) such that all relevant prior knowledge can be immediately accessed to guide action in encountered situations. For example, with the superior organization of knowledge, a chess expert can rapidly perceive a promising move, or a medical expert can rapidly notice an inconsistency in a suggested diagnosis.
Efficient and reliable storage of relevant information in memory is especially important to experts when they engage in planning and complex reasoning that mediate their superior performance. During planning experts have to mentally compare many alternative sequences of actions which creates a great deal of information in working memory. Consequently, beginning chess players do not generate long plans, and it takes years of chess study before chess experts are able to plan long sequences of future moves reliably (Charness, 1989). Chess masters eventually improve their memory skills for planning so much that they are even able to play chess without seeing the chessboard (blindfold chess), thus having to represent the locations of all the pieces on the board during the entire game in their mind. Analyses of the superior ability to plan suggest that experts acquire memory skills, which allow them to rely on long-term memory for storage of generated information (Ericsson and Kintsch, 1995). Recent research on expertise is making it increasingly clear that the vast knowledge of experts has to be well organized and supplemented with special memory skills so as to support memory-demanding planning, design, and reasoning.
Recent research has revealed the complex and intricate structure of expert performance and its associated memory skills. These skills are not attained automatically with experience but require the engagement in deliberate practice, typically designed by teachers. Even the most talented individuals have spent around ten years of intense preparation before attaining an international level of performance in many domains, such as sports, chess, and arts (Ericsson, 1996).
Charness, N. (1989). Expertise in chess and bridge. In D. Klahr and K. Kotovsky, eds., Complex information processing: The impact of Herbert A. Simon, pp. 183–208. Hillsdale, N.J.: Erlbaum.
Chase, W. G., and Ericsson, K. A. (1982). Skill and working memory. In G. H. Bower, ed., The psychology of learning and motivation, vol. 16, pp. 1–58. New York: Academic Press.
Chase, W. G., and Simon, H. A. (1973). The mind's eye in chess. In W. G. Chase, ed., Visual information processing, pp. 215–281. New York: Academic Press.
Chi, M. T. H., Feltovich, P. J., and Glaser, R. (1981). Categorization and representation of physics problems by experts and novices. Cognitive Science 5, 121–152.
Ericsson, K. A. (Ed.) (1996). The road to excellence: The acquisition of expert performance in the arts and sciences, sports, and games. Mahwah, NJ: Erlbaum.
Ericsson, K. A., & Kintsch, W. (1995). Long-term working memory. Psychological Review, 102, 211-245.
Ericsson, K. A., & Lehmann, A. C. (1996). Expert and exceptional performance: Evidence on maximal adaptations on task constraints. Annual Review of Psychology, 47, 273-305.
Ericsson, K. A., Patel, V. L., & Kintsch, W. (2000). How experts' adaptations to representative task demands account for the expertise effect in memory recall: Comment on Vicente and Wang (1998). Psychological Review, 107, 578-592.
Webster's Third New International Dictionary. 1976. Springfield, MA: Merriam
Wilding, J., & Valentine, E. (1997). Superior memory. Hove, East Sussex, UK: Psychology Press
K. Anders Ericsson