Let me distinguish between the record-keeping and constructionist ap-
proaches with a simple example. Suppose an individual—let’s call him Jim—
witnessed a robbery in a convenience store. Let’s say that the burglar was
wearing a black sweatshirt and black jeans, stole money from the cash register,
and stole a radio that was lying on the counter. Suppose that after the burglar
fled, Jim heard a customer claim that the burglar stole a camera. Later on, when
questioned by the police and when testifying in a court of law, Jim must try to
recollect as accurately as possible the details of the crime. For example, Jim
might be asked: ‘‘What was the burglar wearing?’’ or ‘‘What did the burglar
steal?’’
Anyrecord-keepingtheoryclaimsthatwitnessingthecrimecausedJimto
store a new record (or records) in his memory system. When later asked to
recollect the crime, Jim must first search through his memory records until he
finds the record representing the crime, and then try to ‘‘read’’ its contents. If
Jim correctly answers questions about the crime, it is because he was able to
locate the relevant memory record. If Jim forgets, it is because the presence of
so many other memory records made it difficult for him to find the appropriate
memory record or because he was unable to access all the details stored in the
record.
According to constructionist theories, no record-by-record account of past
events is maintained in a storage system. Instead, the cognitive systems for
interpreting and acting on experiences change as a function of the event. For
example, as a result of the crime experience, Jim might learn to avoid conve-
nience stores and to distrust men who wear black clothes. Jim’s cognitive sys-
tems function to anticipate possible future events. When Jim is asked questions
about the crime, he has no memory records to ‘‘read.’’ Instead, he uses the
knowledge currently available in his cognitive systems to derive a plausible
rendition of the past event. For example, he may use his newly acquired dis-
trust of men in black clothes to deduce that the burglar must have worn black
clothes. If Jim forgets, it is because his reconstruction of the past event was in-
accurate. For example, he may remember something about a camera, and so
reconstructthathesawtheburglarstealacamerawhen,infact,theburglar
stole a radio.
The main organizing theme of this chapter, then, is the contrast between
record-keeping and constructionist accounts of memory. A number of cogni-
tive scientists have noted that this contrast is fundamental to understanding
approaches to memory (e.g., Neisser, 1967; Bransford, McCarrell, Franks, &
Nitsch, 1977; Rosenfield, 1988; Howes, 1990). Still, probably no contemporary
theory of memory entirely embodies the record-keeping theory. Even contem-
porary theories that may be characterized as predominantly record-keeping
also make use of constructionist principles (see Bahrick, 1984; or Hall, 1990).
For example, a theory based primarily on record-keeping may claim that peo-
ple resort to reconstructing the past when they fail to find a relevant memory
record. So the record-keeping theory discussed in this chapter serves mainly as
a basis of contrast to help make clear how memory does not work. Examples of
contemporary theories that primarily (but not exclusively) embody record-
keeping principles can be found in Anderson (1983), Anderson and Milson
Memory 313