How Dogs Learn
The science of canine behavior trickles down to pet dogs
in new book for dog trainers
How Dogs Learn by Mary R. Burch PhD and Jon S. Bailey PhD; 1999 Howell Book
House; 175 pages; $19.95 in hardcover
History of dog training / Behavior
principles / Notes
Odd as it may seem to many pet owners, there is heated controversy in the US
over methods of dog training. The battle is framed as a conflict between traditional
methods that punish the dog for misbehaving (often referred to in derogatory
tones as the 'pop and jerk' method) and the 'new, humane method' of rewarding
good behavior with praise and treats.
In their excellent book How Dogs Learn, Mary Burch PhD and Jon Bailey PhD debunk
the black-and-white perception perpetuated by these new age trainers with a
100-year history of dog training, a look at learning theories, and a detailed
explanation of how the various tools and methods work. Theirs is an integrated
approach from a scientific point of view in a book for dog trainers and training
instructors that is unlikely to appeal to John Q. Public Petowner. (No matter;
pet owners will definitely benefit as more and more trainers read the book and
learn to adapt its lessons to individual dogs.)
History of dog training
Some of the current proponents of reward-based training, operant conditioning,
and clickers would have dog owners believe that they are the mid-1990s departure
from harsh training methods of the past, but they overstate their case. Part
One of How Dogs Learn presents the history of training methods through brief
descriptions of scientific research and the methods used by well-known trainers.
From Ivan Pavlov's classical conditioning1 and Edward Thorndike's Law of Effect2
to Skinner's work on operant conditioning and the commercial application of
the theories by Keller Breland, Marian Breland Bailey, and Bob Bailey, they
skim through the highlights of scientific discoveries about animal behavior
and its application to dog training. It is interesting to note that, while some
trainers hail operant conditioning in general and clickers in particular as
the 'new wave' of dog training, B.F. Skinner popularized operant conditioning
in his 1938 book The Behavior of Organisms and in 1951 described the use of
a 'cricket' to train animals in a paper titled "How to Train Animals."
By that time, the cricket - today's clicker - had been in use for several years.
In 1910, years before Skinner's published work, German dog trainer Colonel
Conrad Most wrote Training Dogs, a manual that showed an understanding of operant
conditioning. Most began training police dogs and eventually trained guide dogs
for blind owners. William Koehler, the dog trainer who is most denigrated by
the proponents of operant conditioning, began training military dogs, then segued
into training performance dogs in Hollywood and popularizing dog training for
the general public after World War II. Although Koehler introduced many training
principles that are in use today in modified form, his harsh methods for resolving
behavior problems overshadow his positive contributions in the minds of many
trainers.
Some trainers who followed Koehler adapted his negative-reinforcement methods
to their own style; over the years negative reinforcement evolved into emphasis
on positive reinforcement, and more gentle training methods emerged through
the work of Clarence Pfaffenberger, William Campbell, Ian Dunbar, and Karen
Pryor
Behavior principles
Thus the stage is set by the authors for a discussion of operant conditioning,
a pattern of learning that involves rewarding desired behavior, ignoring or
punishing undesired behavior, analysis of the causes of specific behaviors,
and ultimate understanding of the scientific application of behavior principles
to dog training. In part two of How Dogs Learn, Burch and Bailey cover the basic
principles of behavior in discussions about reinforcement (the reward for an
action), extinction (withholding reinforcement to extinguish a behavior), punishment
(providing negative consequences or eliminating positive consequences of behavior),
and antecedent control (identifying and managing the stimuli that cause a behavior).
Dogs are motivated to perform by the promise of rewards. For some dogs, the
reward is the work; for others, food, play, and praise provide the reinforcement
Behaviors tend to decrease in frequency and eventually stop when rewards are
withheld.
Punishment involves the use of consequences to behavior that are intended to
eliminate the behavior. Punishment must be an immediate consequence of the unwanted
behavior if it is to be effective.
Antecedents are the stimuli that precede the occurrence of the behavior. When
owners put on a jacket, gather car keys, and pick up a leash, Fido wiggles for
joy because he recognizes the preparation for taking a ride.
Part three of the book delves into the diagnosis of behavior. It covers respondent
conditioning (involuntary behaviors related to reflexes and biology such as
Pavlov's dogs salivating at the sound of the bell); analysis of canine behavior
problems; medical and environmental causes of behavior; and additional considerations
for trainers from breed differences to individual dog character traits.
Part four involves teaching old dogs new tricks by using the principles of
operant conditioning, including shaping to establish new behavior; prompting
to set the stage for the behavior; fading to reduce the use of the prompt; chaining
to build behavior sequences; conditioned reinforcement for desired behaviors;
and the use of clickers and target sticks as training aids. The description
of clicker training is easier to understand than most, and the endorsement of
clickers as conditioned reinforcers is not unequivocal. The writers acknowledge
that there are some situations in which clicker training may not be the best
method. Along with noting that some trainers simply prefer traditional methods
of reinforcing desireable behaviors, especially in large group classes, Burch
and Bailey recognize that use of the clicker involves coordination and precision
that many owners find cumbersome.
"Timing is extremely important in clicker training. Trainers who use clickers
incorrectly and make reinforcement errors sometimes shape strange behaviors
and quickly abandon the clicker."3
They also note that: "Clicker training is an excellent method for shaping
new behaviors or improving proficiency. However, it is not the best choice to
reinforce skills the dog has already learned. When the dog has acquired the
desired behavior, trainers are probably better off moving to less intrusive
forms of conditioned reinforcement, such as praise and petting."4
Part five of How Dogs Learn is devoted to decreasing unwanted behaviors through
techniques such as extinction, differential reinforcment, antecedent control,
and punishment. The section on punishment describes several alternatives and
delves into the ethical as well as behavioral consequences of each. The guidelines
for using punishment bear repeating over and over and over:
- Couple punishment for undesireable behaviors with positive reinforcement
for appropriate behaviors.
- Begin with the least aversive method; "If 'no' corrects the problem,
it is not ethical to use anything more intense."5
- If using a punisher such as a shock collar, try it on first.
- Deliver punishment immediately after the infraction.
- Be consistent.
- Teach new behavior to replace the unwanted behavior.
The aim of How Dogs Learn is to make dog training fun for dog and owner and
to minimize the development of unwanted behaviors. We know from studies that
many animals are surrendered to shelters because owners failed to build or maintain
a bond with their pets. An understanding of how dogs learn can make dog training
fun instead of tedious for dog owners. How Dogs Learn will help dog trainers
to help pet owners to build the bond and increase the joy through methods that
make training a partnership.
Notes
- Pavlov's well-known experiments that conditioned dogs to salivate at the
sound of a bell linked a stimulus to an involuntary physical response, not
a voluntary behavior. His work is known as classical or Pavlovian conditioning.
Page 3, How Dogs Learn
- Thorndike's Law of Effect shows that behaviors that are rewarded tend to
increase in frequency. Page 4, How Dogs Learn
- Page 145, How Dogs Learn
- ibid.
- Page 163, How Dogs Learn. This numbered list is a summary of the points
made in this section.
|
