Psychology ~ Behavior ~ 1968
Decades before Brown and Jenkins'
famous publication of the revolutionary autoshaping method, behavioral psychologists had been studying the behavior of rats, pigeons, and the like using the powerful reinforcement tool of hunger.
These pioneering practical psychologists of the past wrote the book on animal
behavior, but traditionally used a method to pre-train the animal subjects to
associate the food dispenser with the food itself. For example, a rat
subject has no unconditioned association between a
metal food dispenser and the actual edible food it dispenses. Behaviorists such
as Skinner, Thorndike and Hull all
experimented with animals that had been pre-trained to associate the food
dispenser (magazine) with the food using a technique called
shaping. The whole process of shaping the organism to associate the
magazine is called magazine training.
Take, for example, a typical scenario of magazine training with
a mouse. Magazine training involves the scientist manually dropping food from
the dispenser while the organism is near a lever that activates the magazine.
Eventually, the animal will learn the function of the magazine and the lever.
Taking advantage of this conditioned connection, the scientist would submit a
hungry mouse to a battery of tests, using the rat's hunger as
an incentive for the rat to complete the tasks. Brown and Jenkins claimed that
magazine training tainted any scientific data gathered by pre-trained organisms.
In a revolutionary study released in 1968, Brown and Jenkins unleashed their
proof that animals could perform tasks involving a magazine without having
to proactively teach the animals that food comes from the food dispenser. They
called this technique autoshaping.
In the above example, the scientist could choose to autoshape
the mouse instead. To do so, the scientist manually drops food from the dispenser
while activating the lever, however doing so at random intervals. Because the
mouse unconditionally desires the food, it will eventually learn that food comes
from the dispenser and that the lever should be pressed to receive food. Although seemingly obvious, the success of autoshaping
challenged established assumptions about animal behavior, specifically that
animals can conditioned themselves without an
intervener.
One minor constraint limits the universality of this popular
technique. Put bluntly, organisms will only autoshape themselves if the
reinforcement is food. In behaviorese: Autoshaping requires the use of a
reinforcement that triggers species specific eating
behavior.
Purdy JE, Markham MR, Bennett SL, Gordon WC, Learning and
Memory, Wadsworth, 2001
Brown, P. L., & Jenkins, H. M. (1968).
Auto-shaping of the pigeon's key-peck.
Journal of the Experimental Analysis of Behavior,11, 1-8.