260 Chapter 32
contradicted earlier psychoanalytic theories that children became attached
because they associated parents with food, the ‘cupboard-love’ theory of
object relations. Bowlby’s view was supported by Harlow’s famous (and
heart-breaking) experiments with infant monkeys. When separated from
their mothers, the infant monkeys spent most of their time clinging to
a cloth-covered model rather than the wire model that they got their
milk from. The cupboard-love theory predicted the opposite, namely that
infants would associate comfort with the wire model that provided them
with food.
Bowlby’s ideas on attachment were greatly influenced by cybernetics,
which considers the ways in which a system can use information in
order to attain its goals. Potential strategies can be illustrated by some
of the different ways a heating system can be designed to maintain a
relatively constant indoor temperature, which is the controlled variable.
A thermostat inside the house provides feedback information on the
controlled variable, boosting the heating when the indoor temperature is
too low. The feedback completes a loop: the heating system influences
the indoor temperature and, by means of the feedback, the indoor tem-
perature influences the heating system. An alternative control strategy
is to use information from predictor variables, sometimes referred to as
feedforward. One instance is a heating system that is turned on in the
autumnandoffinthespring–acrudemethodfavouredbymanyUK
hospitals. A more sophisticated system could use an outdoor thermostat
that turned the heating up when it was cold outside. The information
from predictor variables is not part of a loop: although the season and
the outdoor temperature influence the heating system, the reverse does
not apply. In summary, feedback is from controlled variables whereas
feedforward is from predictor variables.
In order to maintain an optimal security–exploration balance, at-
tachment behaviour is regulated in part by feedforward from predictor
variables. Thus, it makes good evolutionary sense for children to move
closer to a protective adult when they are ill, or when strangers are
around, or when it is dark. This feedforward control is supplemented
by feedback from controlled variables. For example, if you watch young
children exploring in the park, you will see that they usually behave as
though they are attached to their caregiver by a long piece of elastic,
venturing away on their own, but then moving back towards the caregiver
again. The behavioural system that keeps the child from straying too far
can be thought of as a feedback system in which the controlled variable
is the distance between caregiver and child. Thus, too great a distance
between child and caregiver activates the child’s attachment behaviour
and brings the child back to the optimum distance again, in much the same
way that too low an indoor temperature activates a feedback-controlled
heating system and brings the temperature up again.
A wide variety of attachment behaviours can serve the purpose of
keeping the child close enough to a caregiver (just as a wide variety of