by Jeff Stallings, CPDT-KA
I recently reread Temple Grandin’s book “Animals Make Us Human: Creating the Best Life for Animals” in which she brilliantly differentiates between various training techniques, including Positive Reinforcement and Negative Punishment. It got me thinking about how often in dog training these terms are misused and confused, so this post is about explaining these terms and when various techniques might be appropriate.
The most important concept here is semantics because “positive” and “negative” are not in this sense judgments of good or bad but rather indicators of whether something is added (+) or taken away (-). The “reinforcement” part refers to the fact that we are working towards increasing the likelihood that a desired behavior will be repeated.
Conversely, both types of punishment are about decreasing the likelihood of an undesirable behavior. This often-shared graphic sums this up:
Examples of each combination:
An example of Positive Reinforcement is giving (+) a treat the instant your dog performs a desired behavior (sitting on command, for instance), increasing the likelihood that she will do so again in the future. Positive Reinforcement is the highly-effective first course of action for training objective behaviors, such as the sit, stay and down commands. That’s easy to understand.
Negative Punishment refers to taking something away (-) as a means and decreasing the likelihood that a behavior will be repeated. For instance, one method of teaching your dog not to pull on the leash is to stop (-) forward motion when he pulls, only resuming the walk once the leash is loose.
Contemporary training programs use a combination of the above—Positive Reinforcement and Negative Punishment—to teach a dog what’s expected of him.
I also believe that there are (limited) instances in which Positive Punishment is warranted. One common example is teaching a puppy not to bite humans: You calmly but loudly yell “OW!!” (+) when he bites your hand in an effort to eliminate that behavior. The “punishment” in this instance—a loud and unexpected sound—is meant to startle the puppy, providing the opportunity to redirect his mouth to an acceptable chew toy.
Negative Reinforcement means taking something away (pain, for instance) to encourage a behavior (fetching an object.) This example is culled from old-school techniques of yore whereby, in the process of teaching a dog to fetch, his ear is pinched until he picks up the object in question, at which time the pinch is released (-). For training fetch, Positive Reinforcement is by far the better option. In fact, the old-school training techniques that have been most widely criticized and discounted involve Negative Reinforcement.
That said, there are instances in which Negative Reinforcement is appropriate and effective if used judiciously and correctly, and under the guidance of a professional trainer. Example: teaching loose-leash walking by employing a head collar (Gentle Leader or Holt). The head collar applies pressure when the dog pulls that is released (-) the instant he is walking at your side. You have taken something away (pressure) to train a desired behavior (walking by your side instead of pulling.)
Many dog trainers tout the fact that they use only Positive Reinforcement, probably in part because the term Negative Punishment sounds so, well, negative. But using only Positive Reinforcement takes other effective and humane dog training options off the table.
Perhaps someone should come up with a better terms, but in the meantime I will continue to reference Temple Grandin’s analogy to adding (+) and subtracting (-) stimulus to achieve training results.