Logan:  Basic Principles of Learning

            PRINCIPLE: SECONDARY POSITIVE REINFORCEMENT -- HOPE

          If a stimulus more-or-less regularly precedes the occurrence of 
          an emotionally-positive (reinforcing) stimulus, that stimulus 
          will acquire emotionally-positive properties.

     The Principle of Secondary Positive Reinforcement may be viewed as a
special case of the Principle of Classical Conditioning.  If the
unconditional stimulus not only elicits an overt unconditional response
but also produces emotionally positive feelings (rewarding, pleasant,
desirable), these latter feelings also become associated with the
antecedent stimulus.  (From this perspective, the principle is sometimes
referred to as conditioned reinforcement.)  In order to demonstrate that
a stimulus has acquired reinforcing properties, it is necessary to show
that it can function as a reinforcer in the sense that it can be used to
increase the likelihood of responses in substantially the same way that
primary reinforcers do.

     An early demonstration of this type involved the use of chimpanzees
as subjects who were first taught to insert poker chips into vending
machines to obtain various kinds of highly desired foods.  This phase of
the study provided for the pairing of the poker chips with
emotionally-positive stimuli and according to the Principle of Secondary
Reinforcement should have imbued them with hope.  It was then found that
the chimps would willingly solve various kinds of discrimination problems
in order to obtain poker chips that they could later put in the vending
machines for their tidbits of food.  Indeed, different colored chips
could be associated with different subsequent rewards, and became
differentially valued by the chimps.

     Most of the attempted demonstrations of secondary reinforcement have
used stimuli that were initially emotionally-neutral, but this is not
necessary.  A stimulus that is already positive can be increased in
value, and an initially negative stimulus can be converted into a
positive stimulus by being associated with a strong positive primary
reinforcer.  The strength of secondary reinforcement varies with the
types of variables that one would expect from the view that it is an
instance of classical conditioning.  Specifically, the stimulus must
precede the primary reinforcing event, but only by a brief period, and
the more intense or the larger the magnitude of the primary reinforcer,
the greater the secondary reinforcing properties that are transmitted to
the stimulus. 

     The difficulty that arises in attempting to provide an unequivocal
demonstration of the Principle of Secondary Reinforcement is that the
primary reinforcer must be omitted during the second phase so that any
increase in response probability cannot be attributed to the presence of
primary reinforcement.  What this means is that, at the same time that
the secondary reinforcing stimulus is presumably strengthening the
observed response, its acquired reinforcing properties are themselves
undergoing experimental extinction because of the omission of the
unconditional stimulus (the primary reinforcer).  The effect is therefore
necessarily transitory.  Although reasonably durable secondary
reinforcement has been obtained by using procedures designed to increase
persistence, laboratory demonstrations of the principle can best be
described as significant but no overwhelming.

     Even so, the concept of secondary reinforcement has been warmly
embraced by proponents of general reinforcement principles because it
extends those principles beyond the realm of the basic, primary survival
needs.  Very little of our everyday behavior is directly motivated by
hunger or thirst and immediately reinforced with food or water, and
insofar as alternative source of reinforcement is obviously necessary.  In
human society, money has been likened to the poker chips used in the
study with chimpanzees, and has been conceptualized as a GENERALIZED
secondary reinforcer because money is paired with a large number of
different kinds of primary reinforcers.  The contention, which so far has
failed to receive strong experimental support, is that such generalized
secondary reinforcers would indeed have the persistence necessary to
account for typical human behavior.

     Yet another problem is that, in the laboratory, secondary reinforcers
are effective only if the subject is motivated.  The nature of the
motivation can be varied; secondary reinforcers associated with food
reward when the subject is hungry will reinforce behavior when the subject
is thirsty.  But satiated chimps would not work for poker chips to the
same degree that satiated people will work for money.  Which of worse,
reliance on the concept of secondary reinforcement has sometimes been
misleading.  Specifically, it was once presumed that an infant's
attachment to his or her mother derived from the fact that the mother was
frequently associated with the primary reinforcement of feeding.  It is
now known that contact comfort is a primary reinforcing state that is not
dependent on feeding or other basic needs.

     These reservations pertain more to the adequacy of the existing
experimental analyses of secondary reinforcement than to the basic
principle itself.  We strive for praise and social approval, we work not
only for money but for other tokens of accomplishment (such as grades in
school), and many of the rewards that we give ourselves are verbal.  An
important point to remember about secondary reinforcement is that it
depends on our past experiences; not everyone is regarded by the same
things that have acquired reinforcing value to you.

TERMS: Classical conditioning, extinction, hope, motivation, need,
reinforcement (secondary, positive, primary), unconditioned response,
stimulus (antecedent, unconditioned).