This article is from the Training Your Dog FAQ, posted to rec.pets.dogs newsgroup. Maintained by Cindy Tittle Moore with numerous contributions by others.
The principles of classical conditioning were worked out early in this century by Pavlov, and thus is also called Pavlovian conditioning. In the original experiments, a bell was rung, and the subject (as it happens, a dog) was given food; eventually, the dog began to salivate on hearing the bell, apparently anticipating the arrival of the food. This is pure stimulant-response stuff, since the signal (the bell) always comes before the reinforcement, and the dog doesn't do anything to make the bell ring.
So we start with:
1. trainer rings bell (stimulus)
2. dog gets food (reinforcement)
And end up with:
1. trainer rings bell
2. dog drools (response)
3. dog gets food
How can this be used? A great way to use classical conditioning is to teach the dog secondary rewards. Let's say you want to use a particular word or even a particular sound (such as a click) as a reward just because it is simpler than whatever your dog's best primary reward is. So train your dog by saying the word or making the sound and then treating him with a primary reward. He'll start to associate the two quickly and your alternative will become a suitable interim reward for your dog. You'll need to refresh the association from time to time, of course, but it does expand your possible repertoire for telling your dog "You done good!"
If you're observant, you'll also notice that most dogs are classically conditioned. If you say "Sit!" and they sit, that is a stimulus- response sequence no matter how the sit itself was taught.