Who's in charge?
Have you ever felt that you really have no leadership role in your relationship with your dog because your dog is just performing for food? You are not alone.
Consequences have all the power
There is a common belief that dogs, being pack animals, should behave simply because they are subordinate. The problem with this belief is that behavior doesn't work quite so arbitrarily. Behavior doesn't change simply because someone said it should (it might appear to sometimes, but there is always more to it than that). Operant behavior is increased, maintained, reduced, or stopped by consequences which are either reinforcing or punishing.
These consequences can come from great leaders, poor leaders, machines, or, frequently, even by accident—consider this the next time someone tells you "you need to be alpha!"
Envisioning and reinforcing behavior
Great leaders don't allow behavior to simply occur unguided, appearing random and chaotic. They have a vision of which behaviors they want to see and which ones they don't. They see to it that the behaviors they want to see are reinforced and the behaviors they don't want to see are not reinforced.
Clicker trainers focus on positive reinforcement of desired behaviors. Frequently, this positive reinforcement will come through food. Why not? Food is convenient, cheap, quick, and all dogs need to eat anyway.
What is important is that desired behaviors have been reinforced. If a dog performs for food or performs to avoid a correction, that dog has performed for reinforcement. In the first case, for positive reinforcement and in the second case, for negative reinforcement. In neither case did the dog "behave" just because someone said it should.
The clicker and the ABC model
Why do dogs only seem to perform when their trainers have food? Consider the ABC model of operant conditioning (sometimes called "Three Term Contingency"):
Antecedent -> Behavior -> Consequence
If a trainer always has food, then food becomes an antecedent. The trained behavior is less likely to follow without the antecedent of food. The same could be said if a traditional trainer always had a check chain on the dog; the check chain would become an antecedent.
One of the main benefits of clicker training is that the clicker is used as a "bridge" between behavior and reinforcement. This means you can put the food somewhere else and fade it from the antecedents. In time, and if required, you can then put a finished behavior on a schedule of reinforcement and have the behavior performed without always using a reinforcer.
It should also be noted that food is not the only reinforcer available. For more information on fading reinforcers and a short list of common reinforcers, follow these links: