As I mentioned earlier, a big attraction to dog training for me was the science. A desire to understand how things work led me to my job in the Army (radar repair), my first civilian career (industrial equipment repair) and ultimately my current day job (computer programming.) Discovering how dog training works — or more accurately how animals learn — set the hook for my new career.
As I stood in puppy class with Caffeine and the trainer showed us how to lure a sit, a new world literally opened up to me. No forcing, no pulling, no yelling. Show her what you want, get it, reward it.
You can spend a lot of time analyzing dog-dog relationships and dog-human relationships. You can spend hours more arguing over different tools, different techniques, and different models for describing how and why dogs do what they do and who they do it to. You could even create the field of “dog psychology” from whole cloth and get yourself a TV show. But at the end of the day:
- If you want your dog to do something, make it work for them.
- If you want your dog to stop doing something, make it not work for them any more.
- There is no number three. It’s just that simple.
If this sounds oversimplified that’s because it is — but it’s an oversimplification you can do a lot with.
Operant conditioning describes how voluntary behavior is effected by its consequences. When we are teaching our dogs new behaviors or trying to eliminate others, we are employing operant conditioning. (When we are dealing with emotional responses and behaviors that are not entirely voluntary it’s a different story and a different blog post.)
- When a behavior is reinforced the consequence is something that makes it happen more often.
- When a behavior is punished the consequence is something that makes it happen less often.
The term "punished" in the second rule is the source of a lot of confusion. In operant terminology "punish" means reducing the frequency of a behavior. Punishment does not have to be painful or frightening. "Punish" was not the best choice of words, but B.F. Skinner wasn’t thinking about dog trainers and their clients when he picked it.
Alas, the poor word choice didn’t stop there. When describing how punishers and reinforcers are used, we refer to "positive" and "negative", further muddying the waters. Positive is adding something and negative is taking something away. This makes perfect sense right after fifth period algebra, but starts to get sticky later on when you have a leash, some treats in your pocket, and an unruly Border Collie mix to contend with.
- Positive Reinforcement – adding something to make a behavior happen more often.
- Negative Reinforcement – taking something away to make a behavior happen more often.
- Positive Punishment – adding something to make a behavior happen less often.
- Negative Punishment – taking something away to make a behavior happen less often.
Positive reinforcement makes intuitive sense: add something like a piece of chicken or a game of tug to make your dog come to you more often.
Positive punishment seems obvious too: add something like a leash pop or a well-timed reprimand to make a bad behavior happen less often.
Negative reinforcement…how do you take something away to make a behavior happen more frequently? Well, the thing you take away would have to be unpleasant, thereby improving things for the dog. Think about it. Generally not very nice since the unpleasant thing needs to be added first.
Negative punishment makes a lot of sense once you think it through. If the consequence of a behavior makes something good go away, that behavior should stop or at least happen less often. Think about a dog jumping up to take food from your hand and you moving your hand away.
I get confused too. These "four quadrants" are the source of a lot of pointless debate. Some of it is over which quadrants should be used, when, and why. Other debates are over which quadrant(s) are working in what situation. I find these debates maddening enough that I personally leaving the quadrants out of this post.
Before we move on, there’s one important thing I want you to look back at: "more often" and "less often". Reinforcers and punishers affect the frequency of a behavior. They do not necessarily "stop" a behavior or immediately make a behavior guaranteed.
Next week: how we actually use this science.
And since you asked:
Now go find that Gary Numan video. You know you want to.
Photo Credit: Extra Ketchup