Logan:  Basic Principles of Learning

                 PRINCIPLE: PRIMARY NEGATIVE REINFORCEMENT

          If a response is followed by the termination of an 
          innately aversive event, there will result an increase 
          in the probability that that response will recur on 
          future exposures to that aversive event.

     This principle is frequently referred to as ESCAPE LEARNING.  
An organism exposed to a naturally aversive event emits many and varied
responses.  For example, a rat exposed to a mildly painful electric shock
through the grid floor of an experimental apparatus will squeal, prance
around, jump, and probably urinate and defecate.  If in the course of
these activities, the rat happens to jump across a barrier into an 
adjacent safe compartment, that response has received negative
reinforcement and hence is more likely to occur the next time the shock
occurs in the apparatus.  (Note that negative reinforcement is not the
same as punishment, which we will consider separately.  Reinforcement
always increases response probability and, in this case is called
negative only because it is occasioned by a decrease in stimulation.)  As
in this example, the aversive event is usually turned off by the escape
response but negative reinforcement can also be produced by a decrease in
the aversiveness of the event.  Such an incomplete reduction is not as
effective as total termination.

     Acquisition of a response on the basis of negative reinforcement is
typically quite rapid, depending on the intensity of the aversive event.
Generally speaking, escape learning is faster the more noxious the event
that is involved although there is a complication to this rule.  Escape
learning is difficult when the escape response is incompatible with the
innate responses elicited by the aversive event, and these responses may
depend on the intensity of the stimulus.  It is more difficult for a rat
to learn to turn off a strong shock by standing still than if the shock
is weak.

     Escape learning is a model of learning reinforced by drive
reduction.  It is first assumed that a naturally aversive event arouses a
primary pain drive in the organism, which instigates a high level of
activity. If one of the responses that occurs in this process removes the
aversive event, the drive is correspondingly reduced and this outcome
reinforces (strengthens) that response.  It is a principle of biological
survival; drive energizes the organism into action and whatever response
removes the aversive state is reinforced and hence learned.  Some
theorists have used comparable reasoning with respect to the other
primary drives.  For example, we can conceptualize hunger as being an
aversive state and food is reinforcing because it reduces that state.

     As one also encounters in the context of primary positive
reinforcement, there is no way of knowing in advance what events are
aversive to an organism.  Indeed, escape learning is typically used as a
procedure for finding out whether an event is aversive.  That is to
say, we know that an event is aversive to the organism if it will learn a
response to turn it off but some caution is required in using this
circular reasoning.  For example, it is not very surprising that a rat
will learn a response to turn off very bright lights, but a rat will also
learn a response to turn on dim lights in a dark environment.  There are
other contexts in which the optimal level of stimulation is not at zero.
The room temperature of the environment is a case in point because
organisms learn escape responses when it is either too hot or too cold.

     The most dramatic example of too-little stimulation has been
obtained from humans confined in a soundproof room where almost
nothing happens.  Except for periodically being given food, the subject
has nothing to do except lay down and sleep or pace around the room.
After several days in such an environment, the subject will readily
expend effort on any response that will provide some kind of visual or
auditory stimulation and will before long, exercise the escape option of
quitting the experiment.

     At least to some extent, uncomfortable situations are simply
unavoidable in everyday life and we learn to cope with them as best we
can.  Often, we can readily beat a hasty retreat and there is little need
for learning, but many are the homespun remedies for maladies ranging
from a headache to the common cold.  A more challenging task is to
disengage oneself diplomatically from an unpleasant interpersonal
situation such as the married couple who have discovered some unbearable
trait in their spouses.  Finally, the academic drop-out is frequently a
person for whom school, at whatever level including college, has become
unpleasant so that leaving the field appears to be the best escape.

TERMS: Escape learning, reinforcement (primary, negative), escape response.