Reprinted from the Collie Connection, Winter 1999
Well she has one now! OK, it's not perfect, because she still thinks it's optional. But, miracle of miracles, she actually wants to come to me when I call her when we are out on our walks. And she'll come under all sorts of distractions, I dare say more reliably than my other, "normal" dogs trained with traditional methods (...now what does that say?). I cannot fully describe how wonderful it is to see Babe wheel around, when I say "Babe, Come!" and come galloping back to me, fur flying in the wind and with a big Collie grin, planting her front paws on me full force. I have even called her off of running horses, something about which Lisa King can share my utter amazement.
I won't reiterate the various steps we took because I've described them earlier. But eventually, I trusted Babe enough to turn her loose with a light long line, when we went on walks in the country and in our off-leash dog park. The long line was more for training than for security, because the farthest thought from Babe's mind now is leaving our pack (one human, four dogs!). I continued shaping with a clicker, rewarding her for turning to me when I called her. Luckily, the very nature of the recall made it easy to train with a clicker, because if I clicked when she was far away as she turned toward me, she had to come all the way in to me to get her treat--yet the food was not a lure. Also, I taught her to finish the recall by jumping up on me, because I wanted her completely into the "circle of safety" around me. In particular, I did not want her to stop some distance from me and make me lean toward her to give her a treat--that is more threatening and it also throws me off balance if I needed to reach out and catch her.
Now I am starting to add some mild aversives if she does not wheel around as soon as I call her. Finally, Babe takes enough pleasure in coming to me and being with our pack so that there is something to compare the "aversive" to. I use quotes because I'm not talking about a "correction" in which I might act stern or disciplinary with Babe. Rather, I innocently step on the end of her long line and when she hits the end of it and turns around, I act like "what happened???" and then I act happy when she comes toward me.
I am now beginning to train Babe in competition obedience--who knows if Babe will ever enter the competition ring and that's not the point. I am using Babe as a ideal subject to learn the methods described in Morgan Spector's wonderful book, Clicker Training for Obedience. This book has revolutionized Babe's training. He advocates completely embracing this method of positive reinforcement, including an admonishment to train entirely off leash because of our unconscious reflex to use the leash for corrections. There could hardly be a better training philosophy for Babe, who cannot be trained with methods that depend on the use of negative reinforcement or compulsion of any kind.
My experience with Babe's training caused me to become fascinated with two recent columns by Lori Drouin in Front & Finish (January and March 2000), as a counterpoint to Morgan Spector's philosophy. In these columns, Drouin discusses her dislike of clickers, although I can't tell from these columns exactly what it is that she dislikes. From the context of the January column, it seems that she dislikes the purely positive, operant methodology. Her main point seems to be that all dogs eventually need negative reinforcement to be reliably trained for the competition ring. Specifically, she says that dogs learn faster by positive methods than negative ones, but as she puts it, eventually with all dogs you need to "draw a line in the sand." She emphasizes that her students are all pressed for time and do not have the patience, time or diligence required to train successfully with pure operant conditioning, such as Spector advocates in his book. On this point, Spector and Drouin agree in a sense. Spector views negative reinforcement as a way to make up for "trainer error;" he says that the trainer has overreached the dog's capability and needs to make up for that with corrections. His book assumes that you will have the diligence to train with his method and hold off correcting your dog to take short cuts. Terri Arnold makes the same point as Drouin, but differs (I think) from both Drouin and Spector in that she integrates the clicker as one tool in her toolbox--as opposed to saying you must train "all or none," either use clicker-based operant conditioning exclusively or not at all. Drouin, by the way, is nonetheless a big fan of Spector's book.
Babe's training illustrates their points. I tried to short cut training Babe's recall by using negative reinforcement because I feared for her life, but it didn't work. With purely positive methods, Babe's recall training has taken much longer than you could expect to train a normal dog by traditional methods. It took me a year of diligent training, using a well thought-out plan to train Babe to come when called. However, the end result is wonderful--a happy dog with a fast, eager recall and one that is as reliable as my other dogs' recalls trained with traditional methods (which, OK, they still need more work!). And best of all, Babe adores working--she lives to work. She howls and cries constantly when I am training the other dogs and by the time it's her turn, she is frantic to work. Later in her training I added aversives in using mildly unpleasant but emotionally neutral pops to solidify Babe's recall, because it is not reliable enough under distractions. So far, she has tolerated the pops well, but the jury is still out on whether using this negative reinforcement will have the effect that I hope.
Continued in Part 6: purely positive vs. aversive methods in training the stay....
Babe Top|Rescue Top|Site Map