names of the other participants who had attended the last meeting of the sem-
inar (Freeman, Romney, & Freeman, 1987). The subjects were not able to recall
very accurately; about half of their responses were errors. The errors were
revealing, however. Sometimes subjects mistakenly excluded someone who
had attended the last meeting, but usually the excluded person had not regu-
larly attended the seminar. And sometimes subjects mistakenly included some-
one who had missed the last meeting, but usually the included person had
attended most of the other meetings. The errors suggest that the subjects had
extracted the general pattern of attendance from their experiences in the semi-
nar and had used that pattern, reasonably enough, to reconstruct who had
attended the last meeting.
Memory for patterns is also reflected in the tendency for people to remember
the gist but not the details of their experiences. Research has shown that sub-
jects will forget the exact wording of any given sentence in a passage after
reading only a few more sentences, but will usually be able to remember the
meaning of the sentence (Sachs, 1967; for similar results with pictures, see
Gernsbacher, 1985). Research has also shown that after studying a text or a set
of pictures, people will tend to believe mistakenly that a sentence or picture
was explicitly in the set of information they studied, when, in fact, it was only
impliedbytheinformation(e.g.,Bransford,Barclay,&Franks,1972;Harris&
Monaco, 1978; Maki, 1989; Sulin & Dooling, 1974; Thorndyke, 1976). For ex-
ample, if a passage describes an event in which a long-haired customer sat in a
barber’s chair and later left the barbershop with short hair, a subject who had
read that passage may mistakenly believe that the passage also contained a
sentencedescribingthebarbercuttingtheman’shair.Thereasonforthemis-
take is that the implicit information is likely to be consistent with the passage’s
essential themes, which would form the basis of the reconstruction of the de-
tails of the passage.
That memory is better for the patterns or invariants than for the ever-changing
details of experiences is what enables memory to be adaptive, to anticipate the
future. It is the invariants of experience that we are likely to encounter in future
events, so a cognitive system that readily notices such patterns will be better
prepared to respond to new experiences.
Good memory for the patterns or invariants of experience stands in contrast
to our extremely poor memory for the details of the majority of experiences.
Consider—can you describe in detail what you were doing around 3:00p.m.on
May 15th two years ago? Do you remember what the topic of conversation was
when you first met your next-door neighbor? Or what your boss was wearing
when you first met him or her? Or the first 10 sentences of this chapter? You see
the point. What is especially remarkable about our memories is the almost
complete lack of detail they provide about the majority of our past experiences!
And it is easy to demonstrate experimentally that people do not remember very
much about long-past experiences. For example, people have trouble remem-
bering their infant-rearing practices, such as whether they fed their infants on
demand (Robbins, 1963), their formerly held opinions on important political
issues, such as whether they supported busing to equalize education (Goethals
& Reckman, 1973); whether they voted in any given election (Parry & Crossley,
1950); and what they had to eat for dinner six weeks ago (Smith et al., 1991).
Memory 321