Barbara and I hypothesize that the clicker is a conditioned "joy" stimulus that is acquired and recognized through those same primitive pathways, which would help explain why it is so very different from, say, a human word, in its effect. If this is true, another contributing factor to the extraordinary rapidity with which the clicker and clicked behavior can be acquired might be that the click is processed by the CNS much faster than any word can be.
Negative reinforcement can be used to shape behavior. As with positive reinforcement, the reinforcer must be contingent upon the behavior: one must cease "prodding" when the response is correct. Unfortunately, because the prodding, in whatever form results in a change in behavior, the behavior of the person doing the prodding may be positively reinforced, so that, as with punishing, the tendency to lay on with the aversives increases.
The olfactory power of dogs has made headlines again this month, with new research supporting our canine friends' ability to "sniff out" the presence of cancer. While these findings continue to astonish many, one aspect won't surprise fans of operant conditioning: the dogs in this study were clicker trained.
Sure, a collie likes a greasy piece of hot dog as much as the next dog, but it's the game of how to get you to give him that piece of hot dog that he really enjoys. And if the bit of hot dog he expects turns out to be a surprise-a crunchy chunk of dried liver or a game of tug, or even the cue to perform his favorite trick of spinning-wow, the game just got much more interesting. And keeping a collie's interest is the key to training one.
But what if the food-or other reinforcer that is earned, expected, and anticipated-doesn't come right away? What happens when you click for a behavior, touching a target, say, and then delay the presentation of food? What does the animal do? Does it matter? Is there a length of delay that is "too long"?