Latent Learning and Cognitive Maps

194
Chapter 5
Learning
research (e.g., Brennan & Charnetski, 2000). Research is also focusing on how best to minimize learned helplessness and maximize learned optimism in areas such as education, parenting, and psychotherapy (e.g., Jackson, Sellers, & Peterson, 2002).
Latent Learning and Cognitive Maps
Decades ago, Edward Tolman studied cognitive processes in learning by watching rats
try to find their way to food that was waiting for them at the end of a complex maze.
At first, the rats took many wrong turns. Over time, though, they made fewer and fewer
mistakes. The behavioral interpretation of this result was that the rats learned a long
chain of turning responses that were reinforced by food. Tolman disagreed and offered
evidence for a cognitive interpretation.
In one of Tolman’s studies, three groups of rats were placed in the same maze once
a day for several days (Tolman & Honzik, 1930). For Group A, food was placed in the
goal box of the maze on each trial. As shown in Figure 5.14, these rats gradually
improved their performance. Group B also ran the maze once a day, but there was never
any food in the goal box. The animals in Group B continued to make many errors. Neither of these results is surprising.
The third group of rats, Group C, was the critical one. For the first ten days, they
received no reinforcement for running the maze and continued to make many mistakes. On the eleventh day, food was placed in the goal box for the first time. What do
you think happened? On the day after receiving reinforcement, these rats made almost
no mistakes (again, see Figure 5.14). In fact, their performance was as good as that of
the group that had been reinforced every day. The single reinforcement trial on day 11
produced a dramatic change in their performance the next day.
Tolman argued that these results support two conclusions. First, because the rats in
Group C improved their performance the first time they ran the maze after being reinforced, the reinforcement on day 11 could not have significantly affected their learning
of the maze. Rather, the reinforcement simply changed their subsequent performance.
They must have learned the maze earlier as they wandered around making mistakes on
their way to the end of the maze. These rats demonstrated latent learning—learning
that is not evident when it first occurs. (Latent learning occurs in humans, too; for
example, after years of experience in your neighborhood, you could probably tell a visitor that the corner drugstore is closed on Sundays, even if you had never tried to go
there on a Sunday yourself.) Second, the rats’ sudden improvement in performance
FIGURE
5.14
10
Latent Learning
Group C
(reinforced
on day 11)
Mean number of errors
This graph shows the average number of
wrong turns that Tolman’s rats made on
their way to the goal box of a maze.
Notice that when rats in Group C did
not receive food reinforcement, they
continued to make many errors. The day
after first finding food in the goal box,
however, they took almost no wrong
turns! The reinforcement, argued Tolman,
affected only the rats’ performance; they
must have learned the maze earlier,
without reinforcement.
8
Group B
(never reinforced)
6
Group A
(reinforced
on each trial)
4
2
2
4
6
8
10 11 12
Days
14
16