Kapitel 1
The Revolutionary Science of Behavior Change: A Game Anyone Can Play
When Karen Pryor first published "Don't Shoot the Dog" in 1984, few outside professional animal training circles understood the transformative power of positive reinforcement. Today, her once-radical ideas have revolutionized not just pet training but parenting, education, sports coaching, and even business management. The book has become required reading for dog trainers worldwide, earned praise from psychology professors at Harvard, and influenced countless celebrities from Ian Dunbar to Cesar Millan. What makes this slim volume so powerful? It presents a deceptively simple premise: that the science of behavior change is actually a game anyone can play-and one that works across species barriers with remarkable consistency.
Kapitel 2
The Magic of Positive Reinforcement
Imagine being able to transform behaviors without force, threats, or punishment. That's the fundamental promise of positive reinforcement-a concept far more nuanced than simply "rewarding good behavior." A reinforcer is anything that, when occurring with an action, makes that action more likely to happen again. The secret lies in understanding that reinforcers vary by individual and situation. What motivates one person or animal might leave another completely cold.
Take the Wall Street lawyer who transformed his squash game by praising his good shots instead of cursing his mistakes. Within two weeks, he climbed the club ladder rankings and defeated previously unbeatable opponents. Or consider how professional trainers at Sea World use varied reinforcers-fish, stroking, toys, social attention-to keep training interesting and effective for both animals and trainers.
Negative reinforcement, by contrast, involves removing something unpleasant when the desired behavior occurs. When my aunt raises a disapproving eyebrow at my feet on her coffee table, removing my feet makes her face relax, reinforcing my "feet on floor" behavior. Traditional animal training relies heavily on this approach: horses turn when rein pressure ceases, lions back onto pedestals to avoid whips. While effective, negative reinforcement always contains a punishing element that can produce unwanted side effects.
Timing is everything with reinforcement. The closer in time the reinforcer follows the behavior, the more effective it will be. This explains why clicker training works so well-the click precisely marks the exact moment the desired behavior occurs, even if the actual reward comes seconds later. For humans, verbal markers like "Good!" or "Yes!" serve the same purpose.
The ideal reinforcer should be as small as possible while still being effective. Smaller reinforcers allow for more repetitions before satiation occurs. When a zoo keeper gave entire carrots to reward a panda, training progressed slowly because she could only manage three reinforcements in fifteen minutes. Generally, one small mouthful is sufficient-a grain or two for chickens, a quarter-inch meat cube for cats, or even raisins for polar bears.
Occasionally, a "jackpot"-a surprise reward approximately ten times larger than the normal reinforcer-can mark breakthroughs or motivate resistant subjects. At one advertising agency, the president would throw unexpected office parties with caterers, musicians, and champagne, creating tremendous morale boosts. A dolphin named Hou resumed activity after receiving fish "for nothing" during a twenty-minute inactive period. While the psychological mechanism isn't fully understood, jackpots seem to relieve feelings of oppression and resentment.
Kapitel 3
Shaping: The Art of Building Complex Behaviors
Shaping creates behaviors that would never occur by chance-like dogs turning backflips or dolphins jumping through hoops-by reinforcing small tendencies in the right direction and gradually shifting criteria toward an ultimate goal. The scientific term is "successive approximation," and it's a powerful tool for teaching complex skills.
This approach works because behavior naturally varies. No matter how elaborate the final behavior, you can establish intermediate goals and find some existing behavior to use as a starting point. For example, to train a chicken to "dance," you would begin by watching its natural movements and reinforcing any motion that resembles dancing.
We're all familiar with shaping processes in everyday life-from childrearing to learning physical skills like tennis or typing. Our success ultimately depends not on expertise but on persistence, though good shaping techniques can minimize repetition and make every moment of practice count.
The ten laws of shaping provide a roadmap for effective behavior building:
1. Raise criteria in small enough increments that success remains realistic
2. Train one aspect of any behavior at a time
3. Put current response levels on variable reinforcement before raising criteria
4. When introducing new criteria, temporarily relax old ones
5. Plan ahead so you know what to reinforce next if progress comes quickly
6. Maintain consistency in trainers for specific behaviors
7. Change approaches if one isn't working
8. Don't interrupt training sessions unnecessarily
9. Return to basics if performance deteriorates
10. End sessions on a high note while you're still ahead
When raising criteria, stay within the range the subject is already achieving. If your horse clears two-foot jumps sometimes with a foot to spare, you might raise some jumps to two and a half feet-but not all to three feet, and certainly not to three and a half feet. Pushing too far too fast risks breakdown in performance and development of bad habits.
While working on a specific behavior, focus on just one criterion at a time. If training a dolphin to splash, don't withhold reinforcement once for insufficient size and next for wrong direction-one reinforcement cannot convey two pieces of information. Shape first for size until satisfied, then for direction, before finally requiring both criteria together.
When a learner tolerates occasional skipped reinforcers, they'll likely repeat the behavior with more vigor-"Hey! I did it, didn't you see me? Look! I'm doing it again!" This intensified response, called an extinction burst, enables faster progress toward goal behaviors. Skilled trainers may deliberately omit reinforcers to provoke more vigorous responses.
Kapitel 4
The Training Game: Learning Without Words
Even understanding shaping principles requires practice to apply them. Shaping isn't verbal but a nonverbal skill-an interactive behavioral flow like dancing or surfing that must be experienced. The Training Game develops these skills while providing entertainment.
In this game, one person leaves the room while others decide on a behavior for that person to perform-something simple like touching a lamp or sitting in a specific chair. When the subject returns, the group reinforces progress toward the goal behavior using only clicks, claps, or other sounds (no words allowed). The subject must figure out what behavior will earn reinforcement.
What this game demonstrates is not some Machiavellian nature of reinforcement training but the hazards in assuming verbal communication is all-important. The experience of nonverbal learning is especially useful for professionals who instruct others: teachers, therapists, supervisors. Once you've been the "animal," you can empathize with subjects exhibiting behaviors you're shaping but who haven't yet comprehended what they should be doing.
Professional trainers use several techniques to accelerate shaping. In targeting, widely used with sea lions and other performing animals, the trainer shapes the animal to touch its nose to a specific object-a knob on a pole or the trainer's closed fist. By moving the target, the trainer can elicit various behaviors like climbing stairs, jumping, or entering crates.
Mimicry comes naturally to many animals, birds, and people. Young creatures learn by watching and copying their elders' behavior. While psychologists often consider "learning by observation" a sign of intelligence (with primates excelling at it), this skill likely depends more on a species' ecological needs than intelligence alone.
Modeling-physically guiding a subject through desired movements-works best when combined with shaping. Rather than merely pushing the subject through motions repeatedly, effective trainers remain sensitive to the slightest effort to initiate proper movement and reinforce that moment, gradually fading the physical guidance.
Kapitel 5
Stimulus Control: The Secret to Reliable Behavior
Stimuli are anything causing behavioral responses. Some are unconditioned or primary stimuli requiring no training-we naturally flinch at loud noises or follow appetizing smells. Other stimuli become meaningful through learning and association with reinforced behaviors. These learned signals-like traffic lights making us stop and go, or our response to a ringing telephone-guide countless daily behaviors.
Conventional trainers typically start with the cue ("Sit!") before the behavior exists, then physically manipulate the subject into position. After many repetitions, the dog learns to sit to avoid being pushed, making these commands essentially conditioned negative reinforcers.
In contrast, operant conditioning shapes the behavior first, then introduces the cue as a "green light" signaling an opportunity for reinforcement. Why give a command for something the subject can't yet understand? Once the behavior is reliable-like a dog sitting quickly and neatly in various locations-the cue becomes a conditioned positive reinforcer, guaranteeing reinforcement will follow.
Complete stimulus control isn't achieved until the animal learns both to respond to the cue and not to perform the behavior without the cue. Perfect stimulus control requires four conditions:
1. The behavior always occurs immediately when the cue is given
2. The behavior never occurs during training without the cue
3. The behavior never occurs in response to different cues
4. No other behaviors occur in response to this cue
A discriminative stimulus-a learned signal-can be absolutely anything the subject can perceive. Words, flags, lights, touch, vibration, even popping champagne corks will work. Blind dolphins can learn behaviors through touch signals, while sheepdogs respond to hand signals or whistles.
Once a stimulus is learned, it can be gradually reduced or "faded" until barely perceptible while still getting results. This creates seemingly magical performances, like orchestra conductors who establish signals through dramatic gestures, then fade them to subtle shoulder movements.
Behavior chains are sequences where each behavior is reinforced by the opportunity to perform the next behavior, until reaching the final reinforcement. What makes behavior chains work is that each behavior has a reinforcement history and is under stimulus control. The cues can come from a handler, from the environment, or from the previous behavior itself.
The key insight is that chains should be trained backward-start with the last behavior in the chain, ensure it's learned and recognized, then train the next-to-last one, and so on. This approach ensures you're always moving from weakness to strength.
Kapitel 6
Untraining: Eight Methods to Eliminate Unwanted Behavior
When you've mastered establishing new behaviors, the next challenge is eliminating unwanted ones-from kids fighting in the car to barking dogs, furniture-clawing cats, messy roommates, or demanding relatives. There are precisely eight methods for removing unwanted behaviors:
1. "Shoot the animal" - Permanently effective but obviously extreme
2. Punishment - Popular despite rarely working effectively
3. Negative reinforcement - Removing something unpleasant when desired behavior occurs
4. Extinction - Allowing the behavior to disappear naturally
5. Training an incompatible behavior - Especially useful for athletes and pet owners
6. Putting the behavior on cue - Then never giving the cue
7. "Shaping the absence" - Reinforcing anything that isn't the unwanted behavior
8. Changing the motivation - The most fundamental and humane approach
Method 1 always works-you'll definitely never have that behavioral problem with that subject again. Capital punishment, firing employees, divorcing spouses, changing roommates-all are Method 1 solutions. While effective, Method 1 teaches the subject nothing.
Humanity's favorite method is punishment: scolding children, spanking dogs, docking paychecks, fining companies. Yet punishment is a clumsy way of modifying behavior and often doesn't work at all. The troubling pattern with punishment is escalation-when it fails, we don't try something else; we increase the severity.
Punishment rarely works because it occurs after the behavior, sometimes long afterward, so subjects may not connect punishment with their actions. While prompt punishment might stop ongoing behavior, it teaches nothing new-it doesn't show a child how to achieve better grades.
Extinction occurs when behavior dies out due to lack of reinforcement. If a rat trained to press a lever for food suddenly gets no reward, it will press frantically at first, then gradually stop. The behavior extinguishes like a burnt-out candle.
However, ignoring unwanted behavior doesn't always work with humans, as the act of ignoring itself can be a powerful social response. Extinction works best with attention-seeking behaviors like whining, quarreling, teasing, or bullying. When these behaviors produce no results-no reaction from you-they tend to die out.
Kapitel 7
Training Incompatible Behaviors: The Elegant Solution
One elegant approach to eliminating unwanted behavior is training the subject to perform a behavior physically incompatible with the one you don't want. For example, if you dislike dogs begging at the dinner table, instead of banishing them, you can train them to lie down in the doorway during meals. Since a dog can't physically be in two places at once, begging is eliminated.
I witnessed this method's brilliance during an opera rehearsal when the chorus fell out of synchrony with the orchestra. The conductor found an "s" in the lyrics and had the chorus stress it: "The king'sssss coming." This buzzing sound made it impossible for them to rush through the measure too quickly.
My first use of this method was solving a serious dolphin problem at Sea Life Park. We had three types of performers: six small spinner dolphins, a huge female bottlenose named Apo, and a Hawaiian girl who swam with the spinners. Contrary to popular belief, dolphins aren't always friendly, and Apo began harassing the swimmer dangerously, boosting her into the air or slapping her with tail flukes.
Rather than removing our star performer from the show, we trained an incompatible behavior. We taught Apo to press an underwater lever at the pool's edge for fish rewards. During shows, we placed the lever in the pool whenever the swimmer was performing, making it impossible for Apo to simultaneously harass the swimmer and press her lever.
This method works wonderfully for emotional states too. Some activities are totally incompatible with self-pity: dancing, choral singing, or any highly kinetic motor activity like running. You simply cannot engage in them while wallowing in misery simultaneously.
Kapitel 8
Putting Behavior on Cue: The Surprising Solution
This technique is remarkably effective when nothing else works. It's based on a fundamental principle of learning theory: when a behavior is brought under stimulus control-meaning the subject learns to perform it only in response to a specific cue-the behavior tends to extinguish in the absence of that cue.
I discovered this with Makua, a dolphin who would sink to the bottom of the tank to avoid wearing blindfolds. By rewarding him for sinking, then introducing an underwater sound as a cue and reinforcing him only for sinking on cue, he stopped sinking without the cue and accepted blindfolds willingly.
This method works beautifully with noisy children in cars. Just say "Okay, everybody makes as much noise as you possibly can, starting now!" After about thirty seconds of fun chaos, the novelty wears off. Two or three repetitions usually ensure quiet for the rest of the ride.
Deborah Skinner shared a brilliant application using a black/white disk on her door handle to control her dog's whining. When the black side showed, no amount of barking would open the door; when white showed, the dog would be let in.
Kapitel 9
Shaping the Absence and Changing Motivation
The technique of shaping the absence is useful when you don't have anything particular that you wish the subject to do, just that you want them to stop what they're doing. The technical term is DRO (Differential Reinforcement of Other behavior).
Animal psychologist Harry Frank used this when socializing wolf pups, reinforcing with petting and attention anything that was not destroying property. The only non-destructive behavior the pups displayed was lying on the bed, so evenings were passed peacefully with Harry, his wife, and three increasingly large young wolves watching the nightly news together.
I used this method to change my mother's behavior on the telephone. An invalid living in a nursing home, her calls were filled with complaints about pain, loneliness, and lack of money-real problems I was powerless to fix. I began concentrating on my own behavior, letting her complaints extinguish by responding with neutral "Ah" and "Hmm," while enthusiastically reinforcing anything that wasn't a complaint: questions about my children, nursing home news, or discussions about weather, books, or friends. Within two months, the proportion of tears to chat and laughter completely reversed.
Eliminating the motivation for a behavior is often the kindliest and most effective method of all. When addressing puzzling behavioral problems, we should consider possible motivations like hunger, illness, loneliness, or fear. If we can eliminate the underlying cause, we've solved the problem.
Kapitel 10
Clicker Training: The Revolution Spreads
When Don't Shoot the Dog was first published in 1984, applied behavior analysis was still not in general use. Despite thirty years of dolphin training, these techniques hadn't led to widespread applications in other areas.
Clicker training truly began in May 1992 with a panel discussion between trainers and scientists at the Association for Behavior Analysis meetings in San Francisco, followed by a "Don't Shoot the Dog!" seminar for 250 dog trainers. The plastic clickers Gary Wilkes found in a novelty shop made great teaching tools. One seminar led to others, spawning books, videos, and Internet activities that launched the clicker-training movement.
Due to the explosion of clicker training, I began observing more general effects of reinforcement training that I couldn't have imagined earlier. Any creature shaped with positive reinforcers and a marker signal becomes playful, intelligent, curious, and interested in you.
Even a cichlid fish I trained to swim through hoops and follow targets (using a flashlight blink as a marker) became extraordinarily interactive-splashing water to attract attention, touching noses with children through the glass, and threatening visiting dogs by spreading its fins.
Another long-term effect of clicker training is that behavior, once learned, is not forgotten. This retention might be a fundamental difference between positive reinforcers versus aversives and between training with a marker signal versus just primary reinforcers.
Clicker training dramatically accelerates learning across species. Competent clicker trainers accomplish in days what takes months or years with conventional methods. In dog obedience, where traditional training is standardized, the difference is striking. Conventional training typically requires 1-2 years to develop a Novice competitor, and additional years for Open and Utility levels. Now clicker trainers complete all three levels in just over a year.
Kapitel 11
The Ripple Effect: From Animals to Humans
The laws of learning apply to all creatures, including humans. After experiencing clicker training with pets, people naturally begin to generalize their understanding to human interactions. Seminar participants report profound shifts: "I stopped jerking my dogs around-and then I realized what I was still doing to my kids!" Others note transformations in how they manage staff or interact with everyone in their lives.
Teachers, therapists, and parents of children with developmental challenges now use clicker principles in their work. Parents shape appropriate social conversation, eating, dressing and other skills through reinforcement and marker signals. Even young children can apply these principles-like seven-year-old Wylie who calmed his screaming baby brother in the car by reinforcing periods of silence with grins and lollipop licks.
Speech pathologist Sharon Ames solved her twins' three-hour bedtime ordeal in just three days using pennies as reinforcers for each stage of the bedtime process. The first night she clicked and rewarded them frequently, gradually thinning the schedule until the children were going to sleep within twenty minutes.
The technology has found remarkable success in educational settings like Morningside Academy in Seattle, which takes children at least two years behind grade level. Using precision teaching methods that break skills into small steps practiced in short sessions with self-tracking, they guarantee two full grade levels of improvement per year-and have never had to refund tuition.
Public attitudes about behavioral science have evolved considerably. While some still associate Skinner with dystopian mind control, many more people now embrace positive reinforcement. The internet has transformed clicker training into a global phenomenon, with practitioners from Finland to Singapore sharing techniques and success stories.
There's a palpable excitement in this shared communication and experimentation, reminiscent of early adopters in other technological revolutions like flying or radio. For a technology to spread rapidly, it needs three characteristics: it must be easy, have visible benefits to users, and be learnable in small increments. Clicker training perfectly fits these criteria.
The most important impact of reinforcement theory isn't changing specific behaviors but the effect of positive reinforcement itself. Reinforcement provides information about what's working, giving us control over our environment rather than leaving us at its mercy. People enjoy learning through reinforcement not just for rewards but because they gain control over what happens.
A curious corollary is that reinforcement breeds affection in both subject and trainer. The trainer provides life-enhancing events for the subject, and the subject's responses reward the trainer, creating a comradeship. In human interactions, good reinforcement develops family feelings, cements friendships, gives children courage, and teaches them to be skilled reinforcers themselves. It enhances sexual relationships, which are essentially mutual exchanges of positive reinforcers. Two people who excel at reinforcing each other are likely to be a happy pair.
As individuals and as a nation, we should constantly ask: What am I actually reinforcing? Reinforcement creates a process of continual change, give-and-take, and growth. While some see reinforcement as control or manipulation, societal changes must begin with personal changes. Far from being constricting, reinforcement frees us to experience and enhance not the mechanistic aspects of living but the rich diversity of all behavior.
