Skinner and Thorndike
Skinner's Contributions
to Thorndike's Conditioning
Applied to Rat Olympics
Many methods of conditioning have been used and will
continue to be used in the training of the rats. Many of
these methods come from theories and learning methods
developed by Edward Lee Thorndike. However, many of the
ways in which we implement these learning methods are
slightly different than the way they were originally
intended by Thorndike. The reason for this is the
development of these learning methods through other
learned psychologists such as B.F. Skinner. Skinner
altered and added on to Thorndike's original theories
about learning. We have implemented many of Skinner's
ideas into our methods of training the rats.
B.F. Skinner borrowed heavily from Thorndike in his
work. This is reflected in many of the techniques and
methods of thinking used when dealing with the Rat
Olympics. Among these ideas are his views on positive
reinforcement, punishment, shaping, and the continuous
reinforcement schedule.
Skinner added to Thorndike's theories of teaching
when he addressed positive reinforcement vs. punishment.
Skinner attacked the subject in depth, coming up with the
conclusion that while punishment will fix a temporary
problem, it tends to be the wrong way to go about doing
things. His reasoning was that punishment does a good
job of telling the rat what not to do, rather than what
they should do. Skinner placed his emphasis on positive
reinforcement. By providing positive reinforcement, the
rat learns what it has to do. Indeed, that is its job:
a positive reinforcer, either primary or secondary, is
something that, when added to the situation by a certain
response, increases the probability of that response
recurrence (Hergenhahn 1992).
The primary reinforcement method coupled with a
punishment is used in training the rats to run through
the obstacle course. The positive reinforcer is food
waiting for them at the end of the course, while the
punishment is the trainer picking the rat up and starting
him over when he stalls or does an incorrect movement.
Skinner also added to Thorndike's theories of
learning by introducing "shaping", that is, a form of
incremental learning akin to the childhood game of "hot-
cold". In shaping, the rat is incrementally reinforced
to perform an action: at first the rat may be rewarded
for walking across a platform. He may then next time not
get rewarded unless he steps across a small gap. Still
then from there, he may be eventually forced to jump
across a gap to get to the rest of the maze and to his
reward. This can be coupled with Thorndike's principle
of the law of effect, which states that an action that
brings about a satisfying state of affairs will cause the
connection between that action and good things to be
stronger. Thus, the rat will make the connection that
jumping across the gap and the pleasurable stimulus, and
will therefore jump the gap in the future.
Thorndike was very enthusiastic about the need to
repeat actions to strengthen the associations involved
therein. This principle lies with us today, as we all do
homework and go to practice for various sports and
activities. Skinner also believed in the use of a
continuous reinforcement schedule: reinforcing the rat
every time it performed a correct action. Again, the
combination of the two merge and yield a rat that pairs
the action/stimulus with the response of eventually
getting food.
Despite the success of the continuous interval
reinforcement schedule at the outset of training, it does
not produce the best results in the long run. To obtain
optimal performance, it is necessary to start with a
continuous interval and then move to a variable ratio
reinforcement schedule (VR). The VR schedule operates by
reinforcing the rat after a task has been completed a
certain average number of times. It is important to note
that it is an average number; thus, with the average
being 5, the rat could be reinforced the seventh time,
the second time, and so forth as long as they all grouped
around every fifth time. This method is superior to the
continuous reinforcement method because it produces both
a greater resistance to extinction and higher response
rates. This and other reinforcement schedules are
proposed by Skinner, though we have not used them in this
course so far.
Hergenhahn, B.R. Olson, M.H. (1992). An
Introduction To Theories of Learning. New Jersey:
Prentice Hall.