A behavior chain is a string of discrete behaviors that are combined to create a finished ‘chain’. All dog competitions that I know of require that the dog perform behavior chains, either in a predictable manner (ex: obedience) or as directed by the handler (ex: agility). As a result, all successful trainers have learned to create them, even if they weren’t thinking about it in those terms.
Let’s take a closer look at behavior chains. Understanding how they work may help you avoid (or remove) many undesirable behaviors that you may have allowed to creep into your chain.
Here’s an example of a chain for the Retrieve over High Jump (ROH) exercise in obedience:
The behavior chain for a retrieve over high jump includes the discrete behaviors of sit, stay, run forward, jump, pickup, return over the jump, front, hold, release, and finish.
If a dog makes an error within a finished chain, yet is still allowed to complete the chain, then that error is likely to become part of the final exercise if the error is in any way easier or self reinforcing for the dog.
Here’s an example with the ROH described above. If the dog sits, stays, and runs out correctly, but at the point of the dumbbell pick up the dog stops and sniffs for a few seconds before completing the chain, then that momentary sniff may well become part of the chain if the dog found it reinforcing. How the trainer handles this point of failure is critical.
Sniffing is self reinforcing for most dogs. If your dog is allowed to add this additional piece to the behavior chain, and is then allowed to complete the exercise without clear feedback, then sniffing will become part of the chain whenever the dog wishes to sniff. You may not see it on every retrieve, because your dog may not be motivated to sniff on every retrieve, but you will definitely see it again.
So what should you do if your dog adds an undesirable element to a chain?
You must end the exercise any time the dog deviates from the correct chain. This assumes, of course, that your dog was trained to fluency before the chain was put together, and understands exactly how to be correct. (If not, address that first – don’t train behaviors within a chain; break them out and train them separately). If your dog does not know what is right, then telling him what is wrong is likely to lead to shutting down or avoidance behaviors. Try not to go there.
How about behaviors that are not self reinforcing? Those will show up by random chance only (they will neither increase nor decrease), unless the dog decides that you actively WANT the behavior. So…if your dog lifted a paw as he was sent to fetch, then that “paw lift” will only become part of the exercise if the dog either enjoyed the behavior, found it easier than two feet down, or believed that you wanted it (because you somehow reinforced it; possibly simply by continuing to move forward with the exercise) But if that behavior adds nothing to the dog’s enjoyment, is not easier, and the dog doesn’t think that you want it, then that behavior is no more likely than random chance to show up in the future. Since most random behaviors do not become part of a final exercise after one episode, that is rarely a concern. The bigger issue is when the dog found the behavior reinforcing – paw lifts are not reinforcing for most dogs. Of course, if you notice that your dog now frequently raises a paw before being sent on the ROH, it’s likely that your dog added the behavior by random chance, made the association between the behavior and being allowed to complete the chain, and now it is a part of the final chain. While unusual, you’ll see this happen on occasion. How you handle it is up to you – if you want to get rid of the behavior, then break out that piece of the exercise (the stay before the send), and reinforce that piece when the paw stays down. If the paw is raised, re-set the dog and try again. If you don’t care if the behavior shows up on occasion, then leave it alone. In some cases, you might actually WANT the paw raise – maybe you have a dog that taps their feet continuously on active stays and teaching a paw raise prevents it. Then you’d break it out, encourage the paw raise, and then expect it 100% of the time when returned to the chain.
Now that I’ve talked about the ROH example, I’ll let you know that I don’t really want to talk about ROH. I want to talk about heeling. We’ll get to that in part 2.