What's Your Master Plan (part 4 of 6)
This is the fourth post of a 6-part series. If you haven't read the others, here are links to the first, second and third posts. This post discusses the third of 5 meta-lessons that form the basis of the training I've done with my youngest dog Wyrm.
Lesson three: His behavior has consequences.
Sounds kind of dramatic, doesn’t it? “Consequences” sound like something to be avoided. Actually, according to dictionary.com, a consequence is the effect, result, or outcome of something occurring earlier. When a reward follows a behavior, the dog associates the behavior with it’s consequence, the reward. Consequences are how dogs learn. Teaching Wyrm that actions have consequences has nothing to do with punishment. It’s about showing him as often and consistently as possible that his behavior produces consequences, and that he can control his experience of the world by changing his behavior. He can access things he wants by engaging in behaviors I like, and therefore reward. He can also avoid consequences he dislikes by choosing not to engage in behaviors that reliably result in me limiting his access to things he likes.
So what kinds of consequences are we talking about here? Like so many things in life, the answer is “it depends”.
If I’m trying to teach something new, I use high-value rewards like food and (his favorite) a tug toy. Dogs are individuals and their favorite things will vary. If I want him very interested in learning how to do something new, I am going to provide a powerful reward.
In everyday life, using behaviors he has lots of experience in (for him, the recall and down-stay), I use lower-value rewards. This is what everyone seems to want to do with their dogs, but few people are willing to put in the time and effort in it takes to get there. You can’t go directly from continuous food reinforcement to low-value reinforcement, and you absolutely can’t go from continuous food reinforcement to no reinforcement.
Nobody wants their dog to only respond while they are carrying a treat pouch. The way we get to high-performance without high-value rewards is to gradually fade the food rewards, replacing them with lower value rewards, like praise and affection. Eventually, we differentially reinforce, giving high-value rewards only for high quality performance and low-value rewards for lower-quality or adequate performance. Over the course of several months, the behavior improves to the point that what was once a high-quality performance is now a typical, adequate performance. At that point the performance is almost always high quality, even though the high value rewards are rare. Now we just need to maintain the quality of the behavior, which we do by consistently reinforcing it with low-value rewards.
I need to stress the importance of consistent reinforcement. A dog will work hard to learn a new skill for a high value reward. Once that skill is mastered, it can be maintained with low value rewards, but only if they are consistent. Remember that your dog does things for his benefit, and if engaging in the behavior no longer benefits him, it will fade and extinguish.
The most common low-value reward I offer Wyrm for doing what I ask is a combination of praise and freedom. These rewards work great for Wyrm as the praise lets him know when he’s done something I like, and whatever he is doing when he has freedom is exactly what he wants to be doing, which is a great reward. This is known as the Premack Principle, and it’s an extremely powerful motivation, but it only works if you take the time to teach the behavior, fade the high-value rewards, and consistently give low value rewards.
The way I use the premark principle most often is when we are out walking in the woods, I ask Wyrm for a down-stay or recall often while nothing exciting is happening. When he gives me what I want, I praise him and then tell him to “go”, which is one of his favorite things to do. He does what I ask, and 95% of the time he gets praise and then gets to do whatever he wants - he doesn’t give anything up by doing what I ask, I let him go right back to sniffing, chasing squirrels, running about - whatever he wants. The other 5% of the time, he gets excessive praise, but doesn’t immediately get freedom. These are the times when I’m actually using the behavior as a tool to control him in the presence of other trail users, dogs, or mountain bikers. The 95% makes the 5% possible. If I only practiced these behaviors when he would lose his freedom, he would then associate the punishment of losing freedom with the performing the behaviors - he would feel like giving up freedom was part of the behavior rather than an occasional occurrence. The low-value reward of praise only wouldn’t be “worth it” to him, and his performance would plummet.
Finally, speaking about consequences wouldn’t be complete if I didn’t discuss what the consequences of undesired behavior was. For Wyrm, the sternest consequence I give is a verbal reprimand, which I do as rarely and as mildly as possible. One of the typical traits of Border Collies is a soft demeanor, which meansI need to be careful to be as measured as possible with him. Raising my voice, even to others, is very stressful for him. More commonly, the consequence for undesired behavior is the opposite of what I described above for desired behavior - undesired behavior involves removing freedoms, usually by using his obedience cues. If he regularly engages in a behavior I don’t like, I simply consistently have him perform a down-stay when those behaviors appear. He’d rather be free than have me impose control over him, so he learns to avoid the behaviors that reliably result in me taking his freedom away. It’s also important to remember that just as the most effective rewards vary from dog to dog, the negative consequences do as well.
It is truly astounding how much effect you can have on a dog’s behavior using very subtle feedback if you are consistent and systematic in how you apply consequences. This crazy little dog is such a fantastic teacher.