Taking Back the Words is our series on vocabulary the dog training industry broke — and why we refuse to stop using it. Some trainers throw out a word once it’s been corrupted. We’d rather take back the truth of it. Through mechanism, with a dash of humor and love.
The scruff
It’s the third time the puppy has jumped on the woman next to him, and this time his person has had it. Hand on the scruff. Face down into his face. The voice nobody uses in public.
The room goes quiet the way rooms do when everyone is waiting for the lecture.
Here’s what happens instead. The trainer walks over and says: "Oh my goodness, he got you upset. Let me show you another way to handle that behavior — one that’ll get you the actual results you want."
No lecture. No vocabulary correction. Thirty seconds later there’s a puppy in a pen, a parent who just learned something, and a room that exhaled.
Notice what we didn’t do. We didn’t scruff the parent for scruffing the puppy. Hold onto that. It’s the whole post.
The people who saw it first
Punishment is the word our own side threw out. Somewhere in the last thirty years, "good trainers don’t punish" became the membership card, and anyone who said otherwise got sorted into the other camp — the shock collar, the alpha roll, the man with the calm voice and the box on the dog’s collar.
That sort was built on real science. It was just built on half of it.
A rat lab, 1944. William Estes shocked rats for pressing a lever. The pressing stopped — and when the shock stopped, the pressing came right back, in full, as if the rats had learned nothing except to be afraid. His study became the evidence for nearly everything the field would later say about punishment.
Harvard, 1953. B. F. Skinner defined punishment in his textbook as two different things: presenting something aversive, or taking something good away. Then he spent the chapter on the first one — on Estes’s rats and shock — and issued the verdict on both. Suppression, not learning. Side effects. The word never recovered.
A lab in 1962. Harold Weiner ran the other experiment. He had adults press a button for points; then he charged them a point per press. The pressing dropped. And when the cost was lifted, the old behavior came right back, intact. Nobody was shut down. Nobody had to be coaxed out from under a table.
1965. Harold Leitenberg asked the question the field is still confused by: is losing your turn — time-out from the good stuff — actually aversive? His answer, from the lab: yes, in the technical sense. An animal will work to avoid it — and the studies that followed him found it would keep working to avoid it even when the avoiding cost it food. That’s what the word means in a laboratory. It has nothing to do with pain.
1966. Nathan Azrin and Werner Holz wrote the definitive review of punishment research and defined the word, honestly, as the delivery of a stimulus — shock, noise, a slap. Then they catalogued what delivery produces: escape, avoidance, aggression, a suppression that spreads to everything. That catalogue is accurate. It’s also, by their own definition, a catalogue of pain.
1972. Alan Kazdin took stock of a decade of response cost — points, tokens, privileges taken away, in clinics and classrooms — and found a procedure that reliably reduced behavior and kept the people in the room.
A hotel bar in Vermont, 2013. Bob Bailey — who spent a career training animals for the Navy and chickens for anyone willing to learn — walked past a table of the field’s brightest names to sit with two Minnesota dog trainers until two in the morning. It was the most candid conversation we’ve ever had with anyone in this profession. What he said: everything he taught applied to an animal in the box. Controlled environment, professional timing, every contingency in the trainer’s hands. Take the animal out of the box, and the world starts paying for behavior too. He knew it. Most of the field still pretends otherwise.
Seven witnesses. The first two tested pain and called it punishment. The next four tested subtraction and found something else entirely. The last one told us why it matters outside a lab.
What the science says
Two different things, one word. Here’s the mechanical difference, and it’s the only part you need.
Delivered. Something arrives — pain, a scare, a hand in the face. The body files the arrival, and it files what was standing there when it arrived: the person. That’s classical conditioning doing what it does. The person becomes a signal for the thing, and the suppression Azrin and Holz catalogued follows that signal around the house. Quiet, on credit.
Subtracted. Something leaves — the game, the treat, the room. Nothing arrives, so there’s nothing to file against the person. What the body does instead is frustrate and look around: what pays now? A dog with a better option available takes it. That’s not a scar. That’s a price signal.
And here’s why you can’t skip it. Rewarding the good stuff closes the deal only in the box. Out here, the environment is a competing schedule — the counter pays in chicken, the squirrel pays in chase, the other dog pays in attention — and it pays intermittently, the schedule Ferster and Skinner spent seven hundred pages on in 1957 and the most extinction-proof one there is. "No consequences" isn’t a neutral choice. It leaves the lottery running while you bid against it with a cheese cube.
The variable that decides whether a subtraction teaches or just hurts is when. Before the dog has a fluent alternative, taking something away is removal with nowhere to go — frustration climbs, nothing works, and at enough magnitude you’re on the road we fenced off in the Calm post. After the alternative is fluent, the same subtraction is information: this one doesn’t pay, that one does, and he already has that one in his body.
Our one line, and you can keep it: punishment means the behavior went down. Pain means you paid for it with the relationship.
Two rooms, same mechanism
The dog. Play biting is the one behavior a puppy arrives already owning, because his littermates taught it with the same contingency we use: bite too hard and the game ends. So we run it from day one. Too hard gets "gentle" — a warning, with the response still available to him. Another hard bite or a lunge gets "too bad," and the human gets up and walks away. Ten seconds. Nothing scary, nothing intimidating, no pain, ever. He comes back softer.
Everything else waits. A consequence for a cue is allowed only once the dog performs it fluently — hand signal or word, among distractions, at the distance and the duration that matter for his actual life. Fluent in your kitchen at two feet isn’t fluent at the park at thirty, and a "too bad" out there punishes a response he doesn’t have in that room. That’s the version of punishment that produces the twenty-month phone call: he got consequences early and fluency never, and "he knows better" is the receipt.
In class, the consequence is the room. Jumping on people, playing too rough, escalating — all three are paid for by access to people and dogs, so that’s what leaves. Thirty seconds held by the parent: freedom and play gone, person still in hand. Later, if it’s needed, one minute in the time-out pen: no people, no dogs. It’s remarkably effective, and for a reason that has nothing to do with the clock. The thing removed is the exact thing that was paying.
One rule makes it work. If he gets too upset displayed by vocalizations or thrashing to get out, the parent goes in the pen with him. Play stays gone. The person comes back. That single move is the difference between losing the game and losing your person, and it’s the line the cruelty argument missed — because the only time-outs that argument had ever seen took the person too.
The human. You got a parking ticket. You were annoyed. You parked legally the next morning, and nobody had to check whether you’d stopped eating.
Now the parent from the scruff. Here’s the candid part: we accept some fallout, sometimes — based on the parent in front of us. When we’ve watched someone grab a puppy and get in his face, we know exactly what that dog gets at home if the time-out doesn’t take. So we’ll accept a little more frustration in the dog from a pen than we’d otherwise choose, because the alternative is a dog we meet again at social maturity, and it won’t be good — or a dog somebody else meets, with a shock collar as the solution. The time-out isn’t only the dog’s consequence. It’s the parent’s exit ramp: something to do with the surge that isn’t the harmful thing. You can’t grab a dog and walk away from him at the same time.
The prediction
Someone is going to tell you time-outs are cruel. Someone else is going to tell you he needs a firm hand. They’re using the same word, and neither one has separated what arrives from what leaves.
Skinner defined two things and tested one. Weiner’s subjects lost points and kept their nervous systems. Your puppy lost the game for ten seconds and came back softer.
So here’s the test for any consequence — anyone’s, ours included:
Did something arrive, or did something leave? Did he still have a way to win? Could he come back to the game — and did his person stay findable?
We’re keeping the word. Punishment — the real kind: the behavior went down, and the dog is fine. Precise, and nuanced, because we’re not fitting your dog to a belief. We’re meeting the dog in front of us, and the person holding the leash, knowing how behavior actually works — with the fallout ledger open on the table.
One question to take home: if your dog never once loses the game for the wrong choice, who’s teaching him the price of things? Because the world will — and the world doesn’t send a parent into the pen with him.
Next week, the other anchor: how we set the boundary before it’s ever tested. It’s on the door.
— Jody & your Dog Life Coaches
The stance behind the series: The Go Anywhere Dog® Manifesto. The machinery: Our Methodology. Earlier in the series: Calm · Leader · Naughty · Obedience.
Want help making room for better choices with your own dog? Puppy classes in Eden Prairie, or in-home training across Minneapolis and nearby suburbs.
Sources
- Estes, W. K. "An experimental study of punishment." Psychological Monographs, 57(3), Whole No. 263 (1944).
- Skinner, B. F. Science and Human Behavior. Macmillan, 1953 — ch. 12, "Punishment." Defines punishment as either presenting a negative reinforcer or withdrawing a positive one; the chapter’s evidence is aversive-stimulation work.
- Estes, W. K. & Skinner, B. F. "Some quantitative properties of anxiety." Journal of Experimental Psychology, 29(5), 390–400 (1941) — conditioned suppression: shock paired with a signal.
- Weiner, H. "Some effects of response cost upon human operant behavior." Journal of the Experimental Analysis of Behavior, 5(2), 201–208 (1962).
- Leitenberg, H. "Is time-out from positive reinforcement an aversive event? A review of the experimental evidence." Psychological Bulletin, 64(6), 428–441 (1965). Kaufman, A. & Baron, A. "Suppression of behavior by timeout punishment when suppression results in loss of positive reinforcement." Journal of the Experimental Analysis of Behavior, 11(5), 595–607 (1968) — timeout functions as aversive even when avoiding it costs the animal reinforcement.
- Azrin, N. H. & Holz, W. C. "Punishment." In W. K. Honig (Ed.), Operant Behavior: Areas of Research and Application, 380–447. Appleton-Century-Crofts, 1966.
- Kazdin, A. E. "Response cost: The removal of conditioned reinforcers for therapeutic change." Behavior Therapy, 3(4), 533–546 (1972).
- Ferster, C. B. & Skinner, B. F. Schedules of Reinforcement. Appleton-Century-Crofts, 1957.
- Bob Bailey, in conversation, Vermont, 2013 — the authors’ own record; no published source.
