constructionist theory. Certainly the record-keeping theory has dominated the
history of memory research and seems to reflect the ordinary person’s view of
memory (Loftus & Loftus, 1980).
I have argued that the evidence overall supports the constructionist theory
over the record-keeping theory. In the second section, I discussed how experi-
ences are retained in memory. Evidence consistent with constructionist theory
is that memory is good for invariants or patterns that endure across many
experiences, but is poor for the details of specific experiences. Usually people
remember the details of a particular experience because those details are un-
usual or distinctive in some way. Even people with very remarkable memory
for details, such as Luria’sS, make use of mnemonic devices and learning
strategies that help them make information more distinctive.
The constructionist approach claims that memory reflects the strength of
connections among elements of the cognitive systems used to perceive, think
about, and act on events. Such connections undergo continuous reconfiguration
in response to experiences. In a sense, memory is only a byproduct of connec-
tions among the components of various cognitive systems. There is no separate
memory system in which information is ‘‘stored.’’ Consistent with the idea of
memory as a byproduct is the assimilation principle: How well people remem-
ber new information about a topic depends on how much they already know
about that topic. Also consistent is the observation that individual differences
in memory are largely attributable to expertise in the relevant domain of
knowledge. General intellectual skill, or skill at memorizing, does not seem to
predict memory for new information from some domain of knowledge as well
as does expertise in that domain.
Especially telling for the constructionist theory is that conscious recollection
of the past depends on current knowledge and on recollection strategies. As
I discussed in the third section, a person’s recollections of the past are often
distorted by misleading questions or general knowledge. For example, eye-
witnesses to crimes and accidents sometimes mistakenly remember details, like
a car going through a stop sign, that they never observed. Usually such mis-
takes are made when someone or some process implies that the details were a
part of the crime or accident. Especially difficult for the record-keeping theory
is the finding that people are often as confident of inaccurate as of accurate
reconstructions of past events.
Although forgetting is common, people certainly are able to reconstruct ac-
curately some of their past experiences. Memory is more accurate when there
is considerable similarity between the retrieval and original learning environ-
ment, a phenomenon called the overlap (also called the encoding specificity)
principle. The constructionist theory explains the overlap principle by claiming
that memory is improved when the retrieval environment activates the same
portions of the cognitive system used to interpret the original environment. For
instance, people are more likely to use information previously learned in one
environment to solve a new problem if the original environment also required
them to use that information to solve problems.
The most important principle of forgetting, called interference, is that the
more information a person must memorize, the more likely the person will
be unable to remember or will be slower at remembering any given piece of
350 R. Kim Guenther