Page 240 - The Social Animal

222 The Social Animal

need to find additional reasons; consequently, they continued to like
the toy.
46
Jonathan Freedman extended our findings and dramatically il-
lustrated the permanence of the phenomenon. He used as his “cru-
cial toy” an extremely attractive battery-powered robot that scurries
around, hurling objects at a child’s enemies. The other toys were
sickly by comparison. Naturally, all of the children preferred the
robot. He then asked them not to play with that toy, threatening
some children with mild punishment and others with severe punish-
ment. Then he left the school and never returned. Several weeks
later, a young woman came to the school to administer some paper-
and-pencil tests to the children. The children were unaware of the
fact that she was working for Freedman or that her presence was in
any way related to the toys or the threats that had occurred earlier.
But it just so happened that she was administering her test in the
same room Freedman had used for his experiment—the room where
the same toys were casually scattered about. After she administered
the test to the children, she asked them to hang around while she
scored it—and suggested, offhandedly, that they might want to
amuse themselves with those toys someone had left in the room.
Freedman’s results are highly consistent with our own. The over-
whelming majority of the children who had been mildly threatened
weeks earlier refused to play with the robot; they played with the
other toys instead. On the other hand, the great majority of the chil-
dren who had been severely threatened did, in fact, play with the
robot. In sum, a severe threat was not effective in inhibiting subse-
quent behavior—but the effect of one mild threat inhibited behavior
as much as 9 weeks later. Again, the power of this phenomenon rests
on the fact that the children did not come to devalue this behavior
(playing with the toy) because an adult told them it was undesirable;
they convinced themselves that it was undesirable. My guess is that this
process may well apply beyond mere toy preference to more basic and
important areas, such as the control of aggression. Partial support for
this guess can be derived from some correlational studies performed
in the area of child development indicating that parents who use se-
vere punishment to stop a child’s aggression tend to have children
who, while not very aggressive at home, display a great deal of aggres-
47
sion at school and at play away from home. This is precisely what we
would expect from the compliance model discussed in Chapter 2.