Understanding Dog Behavior Chains
Going to the door and having your dogs sit when you touch the handle: an example of a desirable behavior chain.
Desirable behavior chains are also formed in canine sports~!
One of the biggest mistakes novice dog trainers make is ignoring the phenomenon of behavior chains. More than ignoring, I think they are simply not aware of the existence of this phenomenon. This is why it's so important for novice trainers to accompany hands-on training with a good portion of their time spent in understanding learning theory. There are several great dog trainer books out there to help novice trainers grasp better the many dynamics taking place when animals learn. Ignoring the existence of behavior chains is a big mistake because it teaches the dog to perform an unwanted behavior before performing the wanted one. Along with the phenomenon of extinction bursts, I feel that behavior chains is one of the most overlooked issues in dog training and behavior modification. But what exactly are behavior chains?
Imagine looking at a chain necklace. We assume a necklace is a whole thing, but at a closer look, we can clearly see that it is a combination of links attaching one to another. In a similar fashion, a behavior can also be a string of behaviors attached to one another. Let's put this example into practice. If you work in an assembly line for a cookie factory, your cookie-making process may be a series of linked behaviors. You may first grab one cookie, fill it with cream, place another cookie on top, and then place the cookie in the box. Rinse and repeat, you may do this behavior chain over and over during the day.
While dogs do not work for cookie factories (even though I am sure they would love to!) many of their behaviors are also linked to each other in a chain. Their behavior chains may not be made of many behaviors, but they potentially could if you train them. Let's think of a common behavior chain in dogs. For instance, what happens when you grab the leash? Your dog's ear prick up at the sound of it, then he runs towards you, then he sits for you so you can attach it to the collar, then he follows you to the door, then he sits before you open the door (if you have trained polite door manners) then you both head outside to enjoy the walk. All these behaviors may happen each day without us even noticing. What keeps that behavior chain alive?The whole chain of behaviors is reinforced by that final, looked-forward-to reward, which is heading outside.
If your dog dreads heading outside though, you may see a totally different chain of behaviors. Scruffy's body may cower at the sound of the leash, he may freeze when he sees you heading his way, and he may bolt away and hide under the chair the moment he sees you are about to snap it on. In this case, the whole chain of behaviors may be kept kept alive from the relief your dog feels when he hides under the chair and you give up on walking him.
Behavior chains are things dogs understand quite well. Indeed, they study our behavior chains as well! Dogs who suffer from separation anxiety, for instance, know well that you are about to leave the house if every day you tie your shoes, grab the car keys, then get your purse and then open the door to get out. Indeed, he may start feeling anxious the moment you tie your shoes.
You will also see many behavior chains in dog competition events. If you ever watch a canine Freestyle event, you will notice how all behaviors are chained together. The dog will perform a play bow, then he will stand up, run in a circle, heel next to the handler, then play dead, etc. In Rally-Obedience, dogs may perform small chains at each stations, for example you may see a stop, down and walk around. In agility, dogs also perform behavior chains as directed by their handlers and the chains may encompass climbing, jumping, going through a tunnel etc.
Behavior chains are much appreciated when the behaviors in the chain are desirable, or at least, tolerated well by us; but, problems start, when, within a chain, a behavior that is not wanted takes part of the chain. Whether you are dealing with behavior issues or compete in a sport with your dog, you need to be aware of this phenomenon and how to tackle links of unwanted behaviors to prevent them from becoming an integral part of the chain. We will see some examples of behavior chains gone wrong in the next paragraph.
Does your dog jump and then sit? You may be dealing with a behavior chain.
Examples of Unwanted Elements in Behavior Chains
Two-behavior chains are quite common in the dog training world. One of the most common unwanted behavior chains is the jumping dog who is confused by the training implemented in the past. This dog will jump on the owner and then sit for his cookie. This behavior chain has been reinforced for so long, that he repeatedly does this even with guests. Behind this case, there is often an owner who has made a mistake or a novice trainer who has instructed the owner to reward desirable, incompatible behaviors regardless of what happens BEFORE. Behind a dog who jumps and sits, there is often a dog who is confused about the training he has received. The dog simply thinks that the unwanted behavior is part of the chain. He has come to the conclusion "OK, every time I jump and then sit I get a reward, so I must continue to jump and sit. Some dogs are more prone to develop behavior chains than others.
Another example is the dog who pulls on the leash and then comes next to the owner in heel position and is given a treat. In this case, the dog develops what trainers call a "yo-yo walk style." In other words, the dog thinks that the correct way to heel is to pull ahead and then come back, pull ahead and then come back, repeatedly. When I was under the apprentice of a trainer, we were watching this dynamic take place in group classes and I remember asking the trainer: "is the dog consciously learning that he must pull ahead and then come back to his owner for the cookie (as seen in behavior chains) or is he not consciously not connecting the two behaviors and is simply going into heel position, being rewarded, and then just thinks to pull ahead because he believes the exercise is over? We talked it with other trainers and concluded that there was no way to really know what was going on in the dog's mind, but that it's was more worthy spending time solving the issue rather than debating why it happened in the first place.
And then you have the dog who barks, then is quiet and the owners gives a treat. The dog soon learns that in order to get the treat, he'll need to bark and then become quiet, which will obviously increase the barking behavior over time, causing the owners to get frustrated and re-think their training methods.
Finally, worth mentioning is an experiment conducted by Ferster & Skinner, in 1957. Pigeons in the experiment were rewarded for pecking at keys that were lit green. Pecks on keys lit red weren't reinforced. After several sessions, it was noticed how the pigeons started believing that pecking on the red key is what led to a key turning green right afterward, which was rewarded. Basically, a behavior chain was accidentally created; the pigeons started believing that pecking on a red key is what led to the key turning green, thus, generating reinforcement.
Now that we have seen some examples of common behavior chains, let's go to the core of the problem which is preventing them from happening in the first place. And what if a behavior chain has already established? If they were allowed to establish, we'll see some methods on how to resolve the issue.
Front Chaining and Back Chaining
There are two ways behavior chains are taught: front chaining and back chaining. In front chaining, the behaviors are chained in order, so you'll have a dog trained to grab a toy and then drop it in a box. In back chaining, you start with the last behavior and work in reverse. This is often used to train a strong retrieve. Back chaining is very effective because the dog is aware that each behavior is getting more reinforcing and more reinforcing because it's getting closer and closer to the final payoff. Read more about back chaining in this article on "Understanding back chaining" Regardless of the method used, when training behavior chains, you'll have to teach each behaviors individually and then add them up only once they are fluent.
Four Strategies to Break up a Behavior Chain
Method 1: Management
Often, the best way to curb unwanted behaviors is preventing them from occurring in the first place through management techniques. So if you know that your dog will jump when guests come over, find ways to prevent him from jumping in the first place such as keeping him on leash and turning away from the guest the moment your dog gets too excited, or keeping him at a distance where he is better under threshold. Another option is preventing the dog from jumping by intervening quickly before the problem behavior takes place by using a positive interrupter, which can be a smacking sound that the dog has learned to respond to and stop whatever he is doing. This method is useful because it prevents rehearsal of problem behavior, but it can be difficult to be always "a step ahead" of the dog" and "step in" before the dog jumps. Dogs are quite quick in their movements and it's easy for them to catch us unprepared.
Method 2: Breaking the Chain
One of the most common ways to deal with a behavior chain is to let it extinguish by breaking the chain. In other words, we would stop providing reinforcement for the chain, so after some time it extinguishes. So in the case of the dog who jumps and then sits, we would stop rewarding all jumps followed by sits and just reward plain sits with no jumping prior. While this method may work, it's not always practical if we cannot control the environment enough. For instance, you may be picky in not rewarding sits preceded by jumps, but the occasional guests may reward the sit right after the jump sometimes, even through eye contact, talking or petting. All it takes is a few episodes of occasional reinforcement for the chain to come back alive, and thrive, even stronger than before, something that fuels the behavior rather than extinguishing it.
Method 3: Adding more Elements to the Chain
In this case, we are trying to break up the association between one behavior and another by adding more elements to the chain. For instance, if the dog jumps and is used to sitting right afterward, we would ask the dog to leave it, sit, stay and then come. This method was crafted in hopes of breaking the association between jump and sit, but it's not proven to be very effective, for the simple fact that dogs are capable of understanding chains that are longer than two behaviors, so you may be stuck with a dog who still jumps, but then does all the following chained behaviors, for the simple fact that cues can pretty well function as reinforcers that lead to the final reward.
Method 4: Creating a Delay
Anther method is to create a delay in hopes of breaking up the association between the elements. So in the case of the dog who pulls and then comes to heel next to you, try delaying rewarding the heeling for a handful of seconds so the dog forgets about the previous behavior and thinks he is being rewarded for the actual heeling.
Trainers seem to have different preference on what to use, I consulted with several, and turns out each trainer has a favorite approach. Some trainers are fans of using positive interrupters, others prefer working as much as they can in preventing rehearsal of unwanted behaviors, some just let the chain extinguish, and others add delays or insert other behaviors in some cases. I say, whatever works best is the way to go. Only one thing for sure I would never do, that is, adding positive punishment to break the chain. First and foremost, because being a force-free trainer, I am not a fan of this quadrant. And secondly, this would lead to much confusion for the dog which will definitively take away that lovely enthusiasm and joy you should see in dogs who are eager to learn and keep trying.
Alexadry © all rights reserved, do not copy.
BREAKING ACCIDENTAL BEHAVIOR CHAINS by Debra Gayle McKnight May 2010
For further reading
- Conditioned Reinforcers in Dog Training
What's a conditioned reinforcer in dog training and how can it benefit your dog? Learn how to take advantage of conditioned reinforcers for stellar results in dog training and behavior modification.
- Dog Training: How to Use Differential Reinforcement ...
How can you use differential reinforcement of incompatible behavior to train your dog? Learn more about this training method and its advantages.
- A Guide to Dog Behavior Modification Techniques and ...
Say NO to coercion and say YES to rewards. Learn effective modern based dog behavior modification techniques. Mdk572 As the words imply, behavior modification entails modifying a dog's behavior for the purposes of increasing or decreasing wanted and.
- Understanding Extinction Bursts in Dogs
What are dog extinction bursts and how can you use them to your advantage when training your dog or modifying behaviors? There are important rules to follow.