What makes a correction fair?

Zeus's first lesson is where the test is stated most rigorously, because the trainer takes the trouble to invert it. The speeding-ticket argument is run forwards - I do not speed, so I do not get tickets, and I am not afraid of my car - and then backwards, which is what makes it a criterion rather than a reassurance: if you were ticketed while driving under the limit, with no reason given, a few times, you would be afraid to drive at all. So fear is produced by unpredictability, not by penalties. The same lesson gives the two limits, and the order they come in is worth noticing: the first is about the handler's conscience - you should feel okay doing it, you should not feel gross - and it is stated before any equipment is mentioned. The second is injury, with the instruction that there are plenty of ways to punish that carry no risk of it, so use those. His preference for the e-collar is derived from the second limit rather than asserted, and he backs it with something checkable: he has held it at maximum on himself. The second lesson compresses the whole thing into one clause - with clarity, there is no fear.

Roxy's three lessons take the test and turn it into something a household can operate. The first states the condition itself, with the speeding tickets and with the crucial addition that a corrected dog is not obeying out of fear - he has other options and that one is no good. The second lesson adds the dimension nobody else supplies: size. A penalty has a magnitude and a duration, and duration is part of the dose. One sentence of being told off is information; the owner is then asked whether he has ever been shouted at for two or three minutes, and answers that he has, and that it was disturbing. The same lesson contains the corpus's clearest instance of a trainer changing his own practice on a colleague's advice - Christian pointed out that his clients were copying how loudly he trained, so he now trains much more quietly and keeps the volume in reserve for distance. And the handover turns the whole thing into a procedure the family can run without him: three constraints, in order - your own comfort, the risk of injury, and an honest comparison between doing nothing and doing something.

Lola is a nervous dog, and her first lesson contains the upstream half of the argument - the half that makes owners willing to consider any of it. The pressures in a dog's life exist whether or not the handler picks them, so the real question is whether we shape those negative reinforcers or let them be whatever they are going to be. The worked example is the owner's own walk: the dog pulls, the lead is tight the whole way, and at the destination the pressure comes off because the handler is standing there looking at a phone - so pulling has been reinforced, hundreds of times, by somebody who never chose to train it. The same conversation produces the line that draws the ethical boundary, and it is drawn against non-contingent pressure rather than against pressure: pressure they have no way to avoid, because they do not understand why it is happening, is majorly problematic. And a few minutes later the owner reaches the diagnosis herself, out loud - that her role on the walk has become restrainer rather than a source of freedom.

Luna's file supplies the exclusion, and it is the half owners most need: when to do nothing. She is sniffing, nothing was asked of her, and the trainer says on camera that he would not punish that - because while she was sniffing there was no clear command. The counter-case is stated in the same breath: had he said let's go and she ignored it, that is direct disobedience and a different matter. At the handover it is turned into a warning about the family's own instincts - a dog gets up, somebody punishes it, and from the dog's side the thought is what did I do, you didn't tell me to do anything. And the session closes on what the trainer names as the handler's first obligation, which is not consequence but clarity: if you have not shown the dog what behaviour you want, you cannot expect it and you certainly cannot correct its absence.

In the shared Rocko/Kodak owner lesson, the trainer explains why he believes consequences must be understandable and connected to actions the learner can control. A separate Kodak-session explanation adds that the consequence should be proportionate to what is being taught. He describes his intended result as learning the requested behavior rather than becoming afraid of the handler or equipment. These passages state the trainer's rationale and intended outcome. They do not independently establish the dog's internal experience or show that the same result follows for every dog or every use of a tool.

Deuce is ninety-six pounds of adolescent who does not know his own strength, and his file supplies the fairness test owners can apply in the moment, along with the corpus's only stated limit on how distraction work may be manufactured. The test first: if you wound him up, you do not get to correct the jump you produced. The analogy is offered without comment - it would be like saying hey, hold my hand, and then. The same instinct appears twice more with strangers. When a passer-by engages him and Deuce goes over the top, the correction is deliberately withheld, because the man started it, and the failure is logged as proofing still to do. And a couple of days later the trainer states a position on method rather than a technique, while walking towards a group of people in a park: he does not believe in using people as a distraction. The people who happen to be there are fair; staging them is not.

Bo's file sets both the ceiling and the floor, in two different vocabularies ten days apart, and then complicates its own rule. To the camera, the fork is stated cleanly: a choice you could live with but want different gets fixed with obedience; a choice that is dangerous and must never happen again - trying to attack another dog is the named example - gets punished, made a big issue of, and the message made unmistakable. To the owner, the same fork arrives in technical terms: pressure is designed to make a behaviour more likely and comes off the moment he complies; a penalty is designed to make one less likely, has to be immediate rather than five minutes later, and has to be obviously connected to what he did. The complication is a session about crate aggression, where the coached sound for a growl is deliberately not authoritative but disappointed - dude, why are you doing that - delivered while the hand stays exactly where it is, because an authoritative correction would be the fight the dog is expecting.

King supplies the part of fairness that keeps owners following the rules at all: credit for time served. Forty-five minutes on place and he stands up to stretch is not the same event as breaking after two minutes, and it is treated differently - he gets it, and a re-send a few minutes later, without anybody getting ramped up. Two minutes in, the collar and back to the spot. The phrase used for the level is that it is totally dependent on the situation, which is an honest description of a judgement rather than a rule. The passage is repeated in three separate published edits of the same visit, which is worth knowing when counting how often it was said.

Sage's go-home lesson supplies the two routes out of an unwanted behaviour and makes the choice between them explicit, which nothing else in the corpus does. It depends on the result you want. Either you differentially reinforce something else - say no, ask for another behaviour, pay that, and everybody is fine again - or you penalise the thing straightforwardly and then reconcile afterwards. The reconciliation is specified rather than implied: after she has got the point, okay, good girl, good job, and move on. The same file carries the rule the family had already been given at an earlier, unfilmed lesson, and it is the one they had broken: you did not tell her to do anything, so correcting her for it will only confuse her. And it names the escalation trigger - if she goes back to the activity soon after being corrected, the next one has to be backed rather than repeated.

Won't correcting my dog make him afraid of me?

The trainers' answer is that it depends entirely on whether he can see it coming and avoid it. The illustration is always the same and it is always on themselves: I got speeding tickets when I was younger, so I stopped speeding, and I am not afraid to get in the car. Then comes the half that makes it a real test rather than a reassurance - imagine being pulled over and ticketed while driving under the limit, with no reason given, a few times. You would be afraid to drive at all. So what frightens an animal is not consequence, it is unpredictability. A dog who can trace what happened to something he did, and who knows what to do instead, is described as coming out of it unbothered: not afraid of the handler, not afraid of the collar, just less likely to make that choice. If your dog is frightened of you, this method's own account says the problem is in the pattern of consequences he has been getting, not in the existence of them.

How do I know whether a correction I'm about to give is fair?

Ask two questions before you do it and three questions about how. The two: does he know what he did, and does he know what to do instead. If either answer is no, what he needs is clarity, not a consequence - and one trainer names clarity as the handler's first obligation, because expecting a behaviour you have never actually shown the dog is the commonest way owners go wrong. The three about how come from a handover, in the owner's own terms rather than the trainer's. Does doing it make you feel awful? Then do not. Could it injure the dog? Then definitely do not. And, honestly, what happens if you do nothing compared with doing something? That last one is meant to be answered rather than assumed in either direction.

When should I do nothing?

More often than most owners think, and the corpus is specific about it. If you did not give a command, there is nothing to correct: a dog sniffing when nothing was asked of her is free, and a trainer says so on camera in the moment it happens. A dog who gets up off a bed you never sent him to has not broken anything - and from his side the thought is what did I do, you did not tell me to do anything. If you wound him up yourself, the excitement that follows is yours rather than his. And if a stranger started the interaction, the dog's failure is something to work on later, not something to punish now. The rule that covers all four is that a correction has to answer something the dog chose to do against something he was asked.

How big should a correction be?

Commensurate with the thing you are teaching, and no longer than it needs to be. The trainers are much more specific about the ceiling than the floor. A penalty has a size and it has a duration, and duration counts: one sentence of being told off is information, and being shouted at for two or three minutes is disturbing and teaches nothing - an owner in one lesson is asked whether that has ever happened to him and says yes, and that it was. There is also more than one dial. Penalty can be increased socially - presence, posture, tone - rather than physically, and those are different tools. The severity itself is set by what the behaviour was: a choice you could live with but want different is fixed with obedience, while something dangerous that must never recur is answered unmistakably. Where an e-collar's starting level should sit is a separate question again, with two different techniques and two rationales behind it, decided by the owner; see the page on that rather than settling it here.

But isn't using pressure on a dog wrong in principle?

The argument the trainers make is that the pressure is already there and the only question is who is authoring it. A dog pulls on the lead; the lead is tight the whole way; at the destination the pressure comes off because the handler stops walking and looks at a phone. That sequence has taught the dog, hundreds of times, that pulling is what makes the pressure stop - and nobody chose it. So the choice is not between pressure and no pressure, it is between shaping those contingencies and letting them be whatever they happen to be. What the same trainer names as genuinely harmful is non-contingent pressure: pressure the dog has no way to avoid because he cannot work out why it is happening. He also says, in the same breath, that doing nothing at all is worse than that.

What do I do after I've corrected him?

Bring him back to normal before you ask for anything else. One go-home lesson names two routes out of an unwanted behaviour and says the choice depends on the result you want: either redirect - say no, ask for something else, pay that, and everyone is fine again - or penalise the thing directly and then reconcile afterwards. The reconciliation is specified rather than left to instinct: once she has got the point, okay, good girl, good job, and move on. The escalation trigger is named too - if she goes straight back to the behaviour soon after, the next answer has to be backed rather than repeated. And the whole thing only counts if you saw it land: if nothing changed in the dog, no punishment happened.

Correcting something the dog was never asked not to do - a sniff, a stand-up, a jump nobody had commanded against.

Winding the dog up and then correcting the excitement you produced.

Letting the telling-off run on. A sentence teaches; two or three minutes of it does not, and the owner is invited to check that against his own memory of being shouted at.

Correcting late - five minutes after the event, when the dog can no longer connect the two.

Training the pull without meaning to - walking with a tight lead the whole way and letting the pressure come off at the destination while you look at your phone.

Expecting perfect behaviour all day without ever having shown the dog what the wanted behaviour is, and then correcting its absence.

Treating every break the same. A dog who has held a command for forty-five minutes and stands to stretch has not committed the same offence as one who breaks after two.

You find yourself telling the dog off for things you never asked him not to do.

The telling-off goes on - you are still at it a minute later, and you are not sure it is landing.

You correct him when you get home, or when you find what he did, rather than when he does it.

Your walk is a tight lead from the door to the destination, and you have stopped noticing.

You get him excited - the greeting, the play, the doorway - and then tell him off for the jumping that follows.

You feel bad after you correct him, and you cannot say why.