There's an old lady I know who has been training puppies for years. And she does a very good job of it. Funny thing, though, she routinely uses positive punishment as part of her approach. Always has, always will. She's completely unaware of learning theories, has never attended a conference or seminar, and never once read a book about dogs. Yet the puppies she's trained are happy, respectful, and well mannered, and calmly attentive to her subtlest gesture.
How is it that she uses positive punishment to such good effect, without creating desensitization or panic or resentment in her pupils? And how does she do it without collar, leash, head halter, treats or clickers? Because this old lady is a dog. This experienced trainer of puppies is my retired brood bitch Carson.
There's a growing tendency among many dog trainers to denounce the use of positive punishment (P+), though properly defined it means only this: "the presentation of an undesirable consequence." For many trainers, P+ is a bit of jargon heavily laden with ugly images of pain, fear and outright cruelty. And there's no denying that historically, dog training has leaned heavily on punitive methodology, much of which is thinly disguised abuse in the name of training. But when we mistakenly equate P+ with abuse, we are ignoring what dogs themselves can tell us about the value of P+.
Whether we like it or not, P+ is quite natural in animal-to-animal interactions. For example, Carson is resting on a sofa. The puppy Ruby approaches, thinking about jumping up to share the space. Hardening her eyes and holding her head very still, Carson growls softly, her message clear: "Leave me alone." Ruby ignores this. Carson escalates the warning to a loud, scary air snap, and the puppy dances back a step but returns almost immediately, clearly thinking this may be a new game. Carson's next move is a threatening lunge that ends in grabbing the offending puppy by the muzzle for a few beats. Now Ruby gets the message and wisely retreats.
As classically defined, Carson's intent in meting out this punishment is to "decrease the likelihood of the behavior in the future." P+ adds something unpleasant as a consequence for a behavior. Being no fool, Ruby learns that jumping on top of her grandmother has unpleasant consequences; an unoccupied sofa is a better choice. P+ is employed by dogs among themselves all the time.
Dire warnings about the effects of using P+ have their basis in fact. Improperly applied, P+ can undermine the relationship, can lead to desensitization, abuse, cruelty, panic. Though these unhappy results are often presented as an inevitable outcome of employing P+, the truth is that these result from the use of poorly applied or extreme P+. If P+ was destructive to relationships, there would be no cohesive pack structure possible among canids. But dogs and wolves do use P+ and still have strong, affectionate, trusting relationships. Clearly then, the problem lies not with P+ itself, but in our application of it. Learning how to effectively use P+ requires that we look to our dogs for clues.
Properly applied with good timing, clarity and appropriateness of scale (something any socially skilled dog does with an ease that leaves human trainers in awe), P+ makes increasing subtlety of gesture possible. Having used P+ to convince Ruby to nap elsewhere, Carson will not have to escalate the punishment the next time. In fact, she will be able to use less to make the same point. As Ruby learns to read Carson more accurately, the stillness coupled with a hardening eye will be sufficient warning. Dog language is built around nuance and subtlety.
But it is in nuance and subtlety where we often fail in our application of P+. We do not start with stillness and shifts in our eyes and breathing. Often, we begin where Carson ended - grabbing for Ruby and "correcting" her. The puppy has no chance to learn that there are subtleties that should be heeded. From the dog's perspective, we rocket from completely unconcerned to furious action without warning ? a scary situation for a dog that can and does erode the relationship. There's no appropriate beginning to our scale of warnings, and no way to back up from there; we are crude, unsophisticated communicators who do not observe proper canine protocol of escalating warnings.
We can become better trainers through careful study of the nuances of how dogs do what they do, and employing the same subtlety in our own communications. Here's an example - Ruby approaches while I'm eating lunch. My first response is to stop chewing, hold my head very still and harden my eyes a bit. If ignored, I then very slowly turn my head toward Ruby, hardening my eyes further. Ignored, I lift a lip ' just a bit' while still pointedly staring at the offender. Ignored, I growl and escalate the lip lift. Still ignored, I growl louder, longer, and finally, lunge toward her with a threat bark and an air snap. This effective and very "canine" approach does not ruin our relationship in any way. Ruby understands this because it matches what other dogs are also teaching her. Like the other dogs, eventually, I can just offer the stillness and hardened eyes as warning. No equipment or treats or clickers needed - just the power of natural interaction shaped in a way the dog can easily understand, without the emotional overtones often present in human/dog interactions.
When Ruby retreats and sits, I quickly switch to offer positive reinforcement for this desirable behavior. Like all social creatures, dogs need two kinds of information from others. They need to know when they are right, and when they are wrong. Though some trainers refuse to give anything but positive responses, nothing in the dog's culture that supports that lopsided approach. A socially sophisticated dog becomes just that because other dogs told him both when he was right ("I'll keep playing with you since you're behaving nicely") and when he was wrong ("don't bump into me again!"). With a clear understanding of what is right and what is wrong, Ruby can make a choice.
Though we may pride ourselves on being positive trainers, we may be surprised to recognize that we do use punishment, however mildly, from time to time. "Timeouts," disengaging from too rough play, or simply withdrawing our attention from a dog - all are punishment, and all may used as part of a humane approach. The use of punishment (+ or -) doesn't mean we are bad trainers. P+ is merely an "undesirable consequence"; we have the full responsibility for deciding how unpleasant that consequence will be. As with any consequence for a behavior, we will be most successful as trainers when we employ P+ appropriately with awareness, and when we can move past our emotional response to the mere mention of punishment.
As dogs and other social animals show us, P+ may be used effectively without destroying the relationship. They also show us that timing, subtlety, appropriateness of scale, and clarity are the critical ingredients in influencing another's behavior. When we learn from dogs, we learn from the best.
"Lessons from the Masters" by Suzanne Clothier