History of
Clicker Training

Thank you very much, Dr. Foxx, for your introduction. I am enormously honored and grateful to ABA and to the behavior analysis community for this award. I am especially honored to be one of such an incredibly distinguished group as (the other award winners) Dr. Baer, Dr. Sato, and Aubrey Daniels.

What the dog trainers are calling clicker training is an application of behavior analysis that was initially invented and developed more than thirty years ago, by Keller Breland, Marian Breland Bailey, and Bob Bailey. It first reached widespread use in the training of marine mammals, which is where I learned it myself. Although the marine mammal trainers use a whistle, not a clicker, their training relied on the same principles and is really exactly the same as what people are now calling clicker training.

Somehow, however, the technology did not transfer from marine mammal trainers to the general public: perhaps because of the focus of the trainers themselves, and perhaps simply because not very many people need to train dolphins.

In 1992, Phil Hineline invited me to speak to this gathering, at that year's annual meeting in San Francisco. I also organized a panel presentation by operant conditioning animal trainers: myself, Gary Wilkes, Gary Priest, and Ingrid Kang Shallenberger. After the conference Gary Wilkes and I put on the first Don't Shoot the Dog! clicker training seminar, for a group of 250 dog trainers in the Bay Area. That's when things really began to pop.

Five years ago, I could not have named a single clicker trainer in the dog-owning community. Not one. Now, I am personally aware of over ten thousand adherents to clicker training. Our company, Sunshine Books, has a website offering a catalog of clicker training books, videos, and gear, as well as news, downloadable teaching materials, articles, and so on. Last month alone (April, 1997), our website had 20,000 hits.

Harper/Collins Basic Books recently published a book by a philosopher, Aaron Lynch, called Thought Contagion, in which he puts forth the premise that ideas spread through the human population like genes, or like viruses. Whether or not one buys the concept, it's a nice metaphor: and clicker training is certainly a splendid example. Thanks to the internet, we can see this happening repeatedly. We get an order from, say, Finland. Then there are two or three orders from Finland; then a dozen. There is a little pocket of infection there, and it is spreading. We can identify these loci of infection not just in the US, Canada, and the UK, but in Australia, Singapore, South Africa, Sweden, Holland this is a planetary phenomenon.

Clicker training differs in a number of ways from standard or traditional dog training, and from lab training as well. First, although we often use food as a primary reinforcer, we use no deprivation. It is simply not necessary. Second, we use no punishment within the shaping. While reprimands might still occur during daily life, we omit punishment and avoid negative reinforcement as much as possible as an instructional tool. Third, the sessions are very brief, perhaps just a few minutes: no drill, no long repetitions; make progress, and move on. Also, we incorporate a LOT of variety: variations in the training, in the tasks, in the reinforcements used, and so on.

One result is that we are seeing very clearly in our dogs the phenomenon of accelerated learning. On our website we maintain an honor roll of clicker trained dogs. We are receiving and posting many observations of behaviors that, with standard training, take months, even YEARS, to establish with any degree of reliability. Now, with clicker training, standard behaviors and repertoires are being accomplished in six weeks. Three weeks. Four days. A big difference.

The biggest difference, however, is the clicker. Not just this mechanical talisman, but what it actually does. The clicker is not just a conditioned reinforcer, a substitute for food. Yes, it is that, and it's also a bridging stimulus, meaning "Food is coming, but later." And it's a termination signal: "Job's done, come collect your treat." And, as Ogden Lindsley has dubbed it, it is an event marker. But it is more specific than that, and much more powerful. I was fortunate to have been able to discuss this question, the functional role of the clicker, with the late Ellie Reese. And with her how shall I put this connivance I've dubbed it an Operant Topography Identifier. I apologize, this is a clumsy term; but then that hasn't stopped us in the past, has it.

The OTI allows you to communicate some very specific concepts to an animal; and it allows the animal to communicate back. Let me give you an example. A beautiful Arabian show horse, a mare, had been brought to a clicker trainer for some remedial training. Show horses are trained to pose with their feet just so, the neck arched, and the ears pricked forward so they look pretty. The standard horse-training way to get the ears up is to swish a whip around the horse's head, which makes the horse look very attentive indeed. This mare, however, had gone past that. When the whip was swished she'd taken to pinning her ears, that is, laying her ears back flat on her neck, and baring her teeth. NOT pretty at all. And of course escalating the swishing just escalated the ugly face.

The new trainer had begun clicking the mare for pricked ears, and my husband, Jon Lindbergh, captured the next event on video. The mare had learned that clicks mean carrots; and that she could make clicks happen. And she had also become aware that the operant had something to do with ears. But what? So she's doing this: flopping her ears this way and that, backwards and forwards, one ear up and one ear down, rotating each individually and then together—quite a show.

Yes, it is charming; but it is also rather sad. We have been training animals for thousands of years, and we almost never ask them to DO something! To bring their own abilities to the table. To think. If you'll excuse the expression.

When an animal participates at this level, the click, the OTI, acquires enormous value. It is "worth" much more than the food. Both the sound and the object become reinforcing. Here's an example of that. A clicker training instructor in Texas teaches disabled people to train their own service dogs. She is wheelchair bound, herself, and her service dog is a Papillon, a little toy breed, about the size of a cat, black and white. The dog is very useful, in spite of being small. It can retrieve pencils or whatever you drop, find the TV remote, pull laundry out of the dryer, etc. Anyway, when they go to training class, on Thursday nights, this little dog gets down off its owner's lap, goes around under the chairs, gets into people's training equipment bags, and STEALS clickers and brings them back. "Here, Mom, can't have too many of these, can we!"

It is not just for the animal that the clicker is valuable. The sound also provides feedback to the trainer. It is essential, of course, that you click EXACTLY the detail of behavior you want. But most of us do not have that kind of timing naturally. And if you use your voice, you can't TELL when your word was late. With the clicker, though, even a complete novice can tell at once that they weren't on time, and thus fix their own behavior. The clicker provides that vital feedback. And this is the KEY, I believe, to the rapid spread of clicker training.

Aaron Lynch quotes communication scientist Everett M. Rodgers on the special case of the spread of a technology. For a technology to spread fast, it has to have these characteristics: It must be easy. It must have visible benefits to the user. It must be something that can be learned in small increments. (I would add that there must be instant results that are reinforcing to the learner.) Clicker training fits the bill. When you see a well-trained standard-trained dog, you tend to say, "That must have taken years, I could never do that." Or, "My dog could never be that smart." In contrast, people see a clicker trained dog in action and exclaim: "How did you do that? Can I do it? Show me. Let me try."

The prepared mind can pick this whole technology up pretty fast. I don't mean the person who is already a behavior analyst, or a dog trainer. I mean the person who is experienced at choosing action vs. discussion, and who can also think analytically. Some top clicker trainers are, in their day jobs, engineers, scientists, computer technologists, lawyers, police, professors, community leaders, or executives of one sort or another. Some of them almost see it at a glance. They speak in terms of "conversion;" they use the word "epiphany." That makes the people who are opposed to or threatened by clicker training pretty annoyed, you'd better believe.

Sometimes whole schools convert. You may not be aware of it, but in most major cities in this country, and some abroad, there are dog training schools that pass 800 or 1,000 dogs and families a month through their programs. They may have fifteen or twenty fulltime instructors on the payroll. When a school of this size converts to clicker training, that doesn't mean just that 1,000 dogs a month are learning new behavior. It means that 1,000 PEOPLE a month are learning something about thinking in terms of reinforcement contingencies. And they go out and teach more. It is not a geometric progression, yet; based on my feedback from the internet. But it is a very steep arithmetic curve.

What does this mean for behavior analysis? Expanding opportunities for opening a practice as a pet behaviorist, obviously. But make no mistake: this is not about dog training. People LEARN to apply our science, with their dogs. But then they see what it means. And they start applying it to their kids: "My God, I've stopped jerking my dog around, but look at what I'm still doing to my kids." They apply it to their students: "I worked with my dogs all summer, and something happened. For the first time ever, I don't have any problem kids in my classroom." They apply it to their employees. And their co-workers.

And it "sticks." Training trainers has always been a great interest of mine. Not training the animals we know how to do that, now—but getting the technology across so people can use it themselves. How do you do that? So these are very exciting times for me. I get wonderful feedback from the internet, where there are a bounty of clicker training discussion lists; I can watch the information flowing…and once people "get" it, they'll tell you: "I can't go back."

However: Ernst Meyr, the great evolutionary biologist, said recently that the difficulty with cultural change is that it is so easily lost. Marian and Bob Bailey and I have often commiserated with each other over the loss of operant technology in some marine mammal facilities. When the technology is transferred verbally over the decades, from trainer to new trainer with no scientific input, it tends to degenerate. All kinds of superstitious behaviors start creeping in, such deprivation, and punishment, and blaming the animals…it slides right back to the Stone Age.

What is going to be needed, to keep this "sushi science" (a reference to Dr. Sato's speech) of behavioral technology growing and becoming part of our culture, is constant fresh input from you. The opportunity and the need are great.

There's another benefit, I think, to the behavior analysis community. Last night Murray Sidman remarked to me that every time a science moves into a new area, new things happen. That is certainly true of clicker training. There are many, many unanswered research questions popping up as this technology spreads wider and wider. The animal that identifies the operant feels elation: social animals give all their innate signals—dogs bark, horses prance, elephants, I am told, run around in circles chirping. What is going on here? What's going on inside the animal? Endorphin release as the reinforcer?

Why do we see "play" behavior associated with the moment when the "lightbulb goes on?" What's going on when the animal becomes conscious of what it's doing, or, for that matter, when the human learns something he or she is NOT conscious of? This stuff is predictable, it's replicable, it's real. Here's another. We used to think that dolphins were smart because they learned by observation; but we are seeing that clicker trained dogs, watching other dogs get clicked, can learn by observation, and learn even quite elaborate repertoires. Let's study it!

Our time is up. Dr. Foxx asked me to tell you that. You're excused. Thank you for your attention. Click!