Don't Shoot the Dog! : The New Art of Teaching and Training

Introduction

"Reinforcement is information, it's information about what you are doing that is working. " Karen Pryor discovered this training dolphins, then realized it applies to everything from teaching children to changing your own habits. The book presents eight methods for modifying behavior, but focuses intensely on one: positive reinforcement with conditioned signals.

Traditional training punishes wrong actions. Pryor's approach marks and rewards right ones. The difference isn't philosophical, it's neurological.

Punishment suppresses behavior temporarily while creating fear and avoidance. Positive reinforcement builds behavior permanently while creating enthusiasm and creativity.

The core mechanism: timing matters more than intensity. A click or word marking the exact instant of desired behavior communicates precisely what worked. Variable reinforcement schedules maintain behavior far longer than constant predictable rewards. This is why slot machines are addictive and why salaries don't motivate as effectively as unexpected bonuses.

Shaping through successive approximation builds complex behaviors from simple ones using ten specific laws. Target training, mimicry, and modeling accelerate the process.

Meanwhile, four negative methods - punishment, negative reinforcement, extinction, and removing the subject - rarely work and often backfire.

The revolutionary claim: any creature shaped with positive reinforcers becomes playful, intelligent, curious, and interested in you. This applies to dogs, dolphins, children, employees, and yourself. Behavior learned this way doesn't require enforcement, it becomes self-sustaining.

The book provides specific techniques for training incompatible behaviors, putting actions on cue, and shaping absence of unwanted behaviors without ever saying "no. ".

Positive vs Negative Reinforcement

Let's start with the foundation. Reinforcement isn't reward, it's information. And the difference between positive and negative reinforcement? It's not about being nice or mean. It's about what you add versus what you remove. Here's the part that matters most. When you reinforce behavior, the timing is the information.

Not the praise, not the treat, not the compliment. The timing tells the learner exactly what worked.

This is why most training fails. You tell your dog to sit. The dog sits. By the time you say good dog, the dog is standing again. What did you just reinforce? Standing up. The same thing happens with people constantly. You tell your kid gee honey you looked great last night.

That's different from saying it at the moment. The delayed version might even backfire. What's the matter, don't I look great now? We trust words to cover our bad timing. They don't. The timing is the message.

At the Bronx Zoo, keepers had trouble with a gorilla that sat in the doorway blocking them from closing it. They waved bananas, tossed in food. The gorilla either ignored them or grabbed the food and ran back to block the door.

A trainer pointed out they were trying to reinforce behavior that hadn't happened yet. That's bribery.

The solution was to ignore the gorilla in the doorway and only give food when it went outside on its own. Problem solved. Reinforcing too early doesn't work. Reinforcing too late reinforces the wrong thing.

This applies to negative reinforcement too. A horse learns to turn left when you pull the left rein, but only if the pulling stops when it turns. The cessation is the reinforcer. Beginning riders keep kicking the whole time, as if kicking is fuel.

The kicking never stops, so it contains no information. This creates the iron-sided horses at riding schools that barely move no matter how often you kick them.

Same with people getting nagged. If the negative reinforcer doesn't stop the instant the desired behavior occurs, it's not reinforcing anything.

It's just noise. The timing gap between behavior and consequence determines what gets learned. Not your intentions. Not your words. The gap.

Review

So here's your experiment. Pick one behavior you want to change - yours, your kid's, your dog's, doesn't matter. For the next three days, catch yourself about to correct or punish. Stop. Ask instead: what do I actually want them to do? Then mark and reinforce only that.

You'll discover something unsettling: most of what we call training is just organized punishment with no information about what works.

Real training feels like play. If it doesn't, you're doing it wrong. The animal that figures things out becomes creative. The one that's corrected becomes careful. Choose which one you're building.