October 21, 2005

Don’t Shoot the Dog

I came across this book in 1989. I had a degree in education and had been teaching at a University for ten years, when I took a new position in another city. Faced with developing new courses and new teaching methods, and suffering a crisis in confidence, I asked a mentor for recommendations of books that might inspire and motivate me. Don’t Shoot the Dog was her suggestion.

A decade later, when I had my first dog that I wanted to train to a competitive level, rather than to just walk beside me and to hop off the sofa when asked, Don’t Shoot the Dog was again recommended to me.

The principle behind Karen Pryor’s reinforcement training is based on behavioural science and nearly a century of controlled experimentation, but the principle in itself is very simple:

  • A reinforcer is anything that, occurring in conjunction with an act, tends to increase the probability that the act will occur again.
A reinforcer increases the behaviour. It is not necessarily a reward. A negative reinforcer is something that increases the behaviour as the subject tries to avoid the reinforcer. Karen use the example of an Aunt who raises her eyebrows as she goes to put her feet on the coffee table, the raised eyebrows are a gentle aversive that increases the likelihood that Karen will keep her feet on the floor.

A punishment is an aversive stimulus that occurs after the behaviour. Punishment does not result in predictable changes to the behaviour.

So a police car on the freeway is a negative reinforcer that increases the chance of your driving within the speed limit. A fine two weeks later in the mail from a hidden speed camera is a punishment that will have no predictable effect on your driving speed. A safe driver program where your eTag was automatically credited with $10 for every kilometre safely travelled uses a reinforcer and would tend to increase the possibility that you would drive within the limit.

No comments: