combination of memory for specific dates and some limited specialized calcu-
lation (Howe & Smith, 1988). The rare calendar calculators whose abilities
extend froma.d.0toa.d.999,999 appear to use a version of known algorithms
that can be mastered by a graduate student within a couple of weeks to reach a
comparable level of performance (Addis & Parson, described in Ericsson &
Faivre, 1988).
A second example of single-subject research to analyze expert performance
draws on the many case studies of memory experts. Studies of expert memory
performances are particularly suited for the laboratory and can capitalize on
the long tradition of experimental research on memory. The same research tra-
dition has primarily used stimuli that have been selected to be meaningless, or
at least has minimized the role of knowledge in order to capture basic memory
processes. It has, however, been difficult to account for vastly superior memory
performance within this tradition, and occasionally investigators have sug-
gested that such exceptional individuals are endowed with structurally differ-
ent memory systems (Luria, 1968; Wechsler, 1952). Analysis of expert memory
performance is difficult even in the information-processing tradition, because it
is virtually impossible to conduct an a priori task analysis specifying the medi-
ating processing steps and the relevant knowledge used to store information
efficiently in memory.
One of the methods available is to use think-aloud and retrospective verbal
reports to identify the knowledge used by an individual memory expert and
experimentally evaluate hypotheses about the mediating role of that knowl-
edge. For each individual expert it is possible to hypothesize which stimuli
could and could not be successfully encoded using the uncovered mediating
knowledge. By comparing memory performances for compatible and incom-
patible stimuli, it is possible to validate hypotheses about the mediating knowl-
edge using the general method developed by Chase and Simon (1979). In a
study of a long-distance runner who acquired an exceptional digit span through
extended training, Chase and Ericsson (1981) found that the runner encoded
sequences of three digits (513) as familiar running times (5 minutes and 13 sec-
onds in a mile race) whenever possible. When the runner was presented with
experimentally prepared sequences of triplets of digits that could not be encoded
as running times (483 would be 4 minutes and 83 seconds), his digit-span per-
formance was dramatically reduced, and for prepared sequences of triplets all
of which could be encoded as running times, his performance was reliably
improved over his performance with random digit sequences. Similarly, Slo-
boda (1991) shows that superior memory performance for classical music by
idiotssavantsismediatedbyknowledgeofthattypeofmusicandcannotbe
generalized to modern atonal music.
Case studies of memory experts have revealed that the knowledge used to
encode the presented information varies greatly from expert to expert. Simi-
larly, the details of the acquired cognitive structures (retrieval structures) to
store information in retrieval form in long-term memory also differ. Chase and
Ericsson (1982; Ericsson, 1985), however, found three principles of skilled mem-
ory that described the general characteristics of essentially all memory experts
who have been systematically studied.
536 K. Anders Ericsson and Jacqui Smith