As I am writing this, trying to decide how to organize a huge backlog of material, Panda is taking a nap. She's curled up in her pen, just a few feet away from where I'm working. We're just back from our morning training walk, an important walk today because we were trying out her new harness. That's one of the many adventures I want to share with you. In January we traveled over to Massachusetts to visit the harness maker who does all the harnesses for the guide dog schools. Panda had two formal fittings and now has her very own harness.

 

This morning we were testing out the new harness. Both of us were getting used to the differences in fit and function. We had been using Ann's old harness which was made for her first guide dog, a Labrador. It fit Panda, sort of, but not nearly as well as this new harness.

 

When we got back home, I gave Panda an opportunity to relieve herself before coming into the house. We now have a solid cue for urination, and a separate cue for defecation. And since she also needs a cue to tell me when she needs a bathroom break, she's learning to ring a bell on the back door when she needs to go out. That's another chapter, and an interesting one on stimulus control. (I might as well warn you now. If you've never house trained a puppy or potty trained a toddler, you may want to skip the last section of this report. It goes into the sort of detail on "potty training" that only animal trainers or doting mothers can appreciate.)

 

Panda knows her daily routine well. After her pit stop she follows me into the house. She usually pauses for a long drink from the cat's water bowl. She has her own bucket outside, but she prefers their bowl. Go figure! The cats think it's disgusting having to share!

 

I have a dog's exercise pen set up in what used to be the dining room. Over the last couple of weeks it has been converted completely into my office/shipping room. The main computers are still in the basement, but I've set up my laptop so I can work upstairs. Panda is handling stairs better and better, but she isn't yet ready to go down into the basement. That's like asking a beginner skier to go down the expert trails, not the best way to build confidence. So I've relocated most of my work upstairs to the dining room.

 

The exercise pen is in a corner, next to the door going into the kitchen. When Panda comes into the house she goes straight into the pen, turns around, backs up into the far corner and waits. I then take off my shoes and coat, and dish out a yogurt cup full of hay cubes for her. At this point Panda is still waiting in her pen. She has learned that waiting is what gets me to give her the hay cubes. I've been building duration, asking her to wait a little longer each time. If she comes forward out of her pen, I simply ask her to back up again. Panda usually stays put. She knows how to operate her human!

 

Panda finished her lunch a few minutes ago, and she's now sacked out taking a nap. She'll be down for another hour or so. Yesterday she had quite the dream. She was whinnying in her sleep, a happy little chortle of a whinny, very cute. Right now she is flat out on her side, snoring ever so softly.

 

When she wakes up, I'll take her out again to relieve, then she'll have another snack and a nap. By then it will be time to run afternoon errands. If I'm not pressed for time, I'll take her with me. Otherwise, I'll leave her at home. Traveling with Panda always requires extra time, not because she's doing anything wrong, but because so many people want to stop and visit with her.

 

Our evenings are spent at the barn where Panda gets to visit with the other horses. We play musical stalls at night, juggling Panda in and out as we work each of the horses. She's learning how to tie, but she isn't ready to stay on a tie all evening. Eventually, as a working guide, she will need to be able to stay on a tie for long periods. For now, like all the other tasks she is learning, we are building the duration slowly.

 

Her favorite place to be tied is in front of the Fengur's stall. Fengur is one of our Icelandics. Before we got Panda he used to look like a small horse. Standing next to her he now looks huge. Fengur likes to poke his head out through the opening for grain in the front of his stall. He looks like a mounted head when he does. It's ever so comical, especially since his great mass of forelock and mane gets shoved back by the bars so he looks like a cross between a mounted moose head and a lion. When Panda is tied next to him, he sticks his head out and grooms her stem to stern. Too cute!

 

At the end of the evening, we generally play "Panda catch". That's the best game of all. "Panda catch" grew out of heel position and Panda's love of play. It's a game she'll eagerly play with just one person, or as many as we have in the barn. I described the origins of it in the first two Panda Project reports. It has now grown into a wonderful game where you can send Panda out and she will boomerang back to you in perfect heel position.

 

Picture this, you're standing at one end of the barn aisle, and I'm at the other with Panda at my side. On my signal Panda moves out away from me and towards you. She trit trots down the length of the aisle to you, circles around behind you and locks in at your side in perfect heel position. That's "Panda catch". It's even more fun out in an arena where she has room to zoom. The best fun is to run off to the side away from her, and then to stop abruptly and stand still with your eyes closed. In seconds she'll be there, pressing in against your leg, telling you she's found you.

 

What interests me in particular about this behavior is how willing Panda is to play the game with anyone and everyone. You don't have to be her special person. She has generalized the exercise so she will lock onto anyone she is with and perform for them. That means that transferring her training from me to Ann should not be a problem.

 

Thinking like a Guide-horse Trainer

"Panda catch" and all the other details of her training grew out of being consistent, and understanding that little things matter. As always principles come first. Most of us learned training the other way around. We learned mechanics first. For the most part we weren't given an organizing framework that put everything together. I think that's the appeal that training methods such as natural horsemanship have for many people. they provide a framework.

 

Clicker training also creates a framework because it is grounded in the principles that underlie all good training methods. It is those principles that I hope to convey with these Panda Project Reports. My goal most definitely is not to teach you how to train guide horses. That's something that is best left to experienced handlers, but I do think there is a very real value in thinking like a guide horse trainer.

 

I've been noticing in my own training the very positive effect this experience with Panda has had on my training. You cannot be sloppy and train a guide. John Lyons has an expression: "The horse doesn't know when it doesn't count so it always has to count."

 

That's true of all training, and especially true in guide work. When I'm with Panda, I stop at curbs, at ALL curbs, even when I'm in a hurry, or thinking about something else. As a working guide ALL curbs will count, so all curbs must count now.

 

When we're out walking and Panda starts to rush forward, I check her speed. It doesn't matter that I might be in a hurry, too. Pulling, especially over uneven or unfamiliar ground, might unbalance her blind handler. Panda doesn't know when it doesn't count, so it always has to count.

 

As you're leading your big horses, try thinking like a guide-horse trainer. Imagine that you are preparing your horse for a blind handler. Does your horse tend to rush ahead of you going into the barn? That might not seem like a problem to you. You can see where the door jam is and the patch of ice that's built up on the sill. But imagine if you couldn't see. Wouldn't it be better if your horse stopped at the entrance way? Suppose you got in the habit yourself of stopping at every doorway, every gate. You could mark each halt with a click and a treat. Stop at every door, click and treat. Stop at the gate going out to the paddock, click and treat. Stop before going into the stall where your horse knows his dinner is waiting, click and treat. It wouldn't be long before you'd see a lot more patience and calm self-control coming into your horse.

 

"Your horse doesn't know when it doesn't count so it always has to count."

 

I've often said that I should hang large banners with phrases like that over my horse's stalls. Everytime you walk down the aisle, you'd see these reminders of the major training principles. That one would certainly get a prime spot in the barn.

 

Another would be: You can not ask for something and expect to get it on a consistent basis unless you have gone through a teaching process to teach it to your horse." That's certainly been a major organizing principle for all of Panda's training.

 

So let me back track now to share some of details of her training and the adventures we've been having.

 

December:

 

The "Panda Pose" and "Leave It" 

After all the traveling Panda and I did together in the fall, December was a quiet month spent mostly at home. It was a month spent mainly consolidating little details in the training, but I did add one major piece, "Panda's pose". That was taught at a clinic I gave at Dolores Arste's. Saturday morning began, as usual in the house, getting organized and discussing the backlog of training questions that had accumulated since the last clinic. Panda knows what to do when she goes into Dolores' living room. She lies down on the carpet and takes a nap. Ann's new guide dog did the same thing. The two lay back to back having a morning snooze. And did anyone have a camera handy. Of course not!

 

When Panda woke up, we were still deep in discussion. Part of what we talking about was the "always there" presence of clicker training. Clicker training for me is not something I do now and then with a horse. I don't have separate clicker training sessions and then go off and do other things with my horses. When I am with my horses, I am clicker training.

 

As the discussion continued, I modeled this concept. I noticed that while Panda was standing still, she sometimes relaxed the muscles around her throat so her head ducked ever so slightly towards her chest. Click and treat, I captured that tiny drop of her chin. Now remember, I use my tongue to make a clicking sound, so I can be very discreet with the training. I can carry on a conversation and at the same time click and reinforce my horse. I don't have to worry about the sharp sound of a clicker distracting anyone's concentration.

 

So we talked training, and I reinforced Panda for tucking her nose towards her chest. I don't think anyone even noticed at first what was happening. They probably thought that I was just reinforcing Panda for standing still. They didn't realize that the click came at the exact moment that I saw that miniscule relaxation of her neck muscles.

 

You get more of what you reinforce. It didn't take long before Panda figured out how to get me to click. She started deliberately flexing her poll and dropping her nose into a beautiful tucked position. By the time we were ready to go out to work the big horses, Panda was offering a perfect "dressage horse" bascule of her head and neck.

 

At lunch I carried this another step further. I built some duration into the behavior. Panda had to hold her pose a little longer. When you go for duration, you will often see your horse trying extra hard to get you to click. Panda didn't just arch her little neck. She rounded her back and shifted her weight onto her hindend. You could put your hand on her withers and feel them lift a couple of inches as she rounded into her pose. Click and major treat!

 

Panda kept showing off her pose and getting better and better at it. She was the perfect demo horse for understanding the shifts that occur as a horse engages it's body. And Panda was like my big horse, Robin. Once she had the pose, she was eager to show it off. I encouraged her by keeping the behavior on a high rate of reinforcement. I know from Robin what a valuable behavior this is. I treat it as a default behavior, meaning I am the cue for it. When Robin sees me, he "poses", and I reinforce him. A horse who is "posing" is practising the very thing I am going to want him to do under saddle. And just as important, he is not practising other things, like pawing to get attention, that I do not want.

 

A horse can't do "nothing", so it makes sense to choose the "something" he's be doing when he's around me.

 

Panda's default behavior has been heel position. That means in the absence of some other specific request from me, when she's with me, this is what she's to do. When we're waiting in line at the post office, or visiting with someone, Panda parks herself at my side in heel position. She isn't fussing or swinging her hindquarters out into somebody's space. She's locked in tight by my side where I know exactly where she is. That's an incredibly useful skill for an animal that goes out in public. And Panda being Panda, she's now one upped herself and added the pose into her heel position.

 

At home during the week I noticed an immediate benefit from this new addition. I had taught Panda lateral flexions (See Report 2) to help develop control of leg speed and to keep her from pulling. Now with the pose in her repertoire, she automatically added that into her leading. She looked like a little carriage horse trotting out in front of me. I loved the feel through the harness, the extra spring in her gaits the added engagement gave her. She felt light in my hand, and so connected. It was a joy to walk with her.

 

This kind of energized trot is self-reinforcing. At this point the need to click and mark it diminishes. The movement itself becomes the reward. Through prior training you've reinforced the horse for each step towards that balance. They understand the process so thoroughly you can now let them enjoy the result.

 

That's why the "click ends the behavior" is such an important concept, especially for performance horses.

 

Every time you click and the horse stops you are marking a step, a layer in the training. That means that these horses really come to understand exactly what earns them reinforcement. They want their treat, so they get themselves organized. That's the key. They get themselves organized, not the rider holding them together between leg and hand, not some mechanical device supporting their head set. The horse learns to take responsibility for his own balance. The result is a freedom and ease of gaits that is beautiful to watch and even more amazing to ride.

 

Of course, little Panda is never going to be ridden, but she will be a working girl. The more correctly she moves, the easier physically it will be both on her and her handler. My client jokes about "guide dog elbow", meaning the strain that comes from holding onto the harness of a dog that is pulling. The reality is the strain is no joke. There is stress on the handler, and the dog. That's something I hope to avoid by putting some of these dressage-horse principles to work in Panda's training.

 

"Leave It"

Each little piece leads to the next. At the Tufts Animal Expo in October Morgan Spector, author of Clicker Training For Obedience, had introduced Panda to the "leave it" game. (See Report 1). He held his closed fist in front of Panda. If she nibbled at his hand, nothing happened, but, if she pulled back from his hand, click, she got a treat. We'd been practicing this simple exercise and expanding upon it so that Panda now understands that drawing back from people who approached her head-on earns reinforcement. I taught her this so that she would have some way of alerting Ann to the fact that somebody was petting her. Since Panda is a people magnet, Ann will need a way to monitor and control access to Panda.

 

Most people understand that an animal in guide harness is working and shouldn't be distracted. They are good about asking if they can pet her. However, we do get some people who come up to her without asking. In the beginning that was a problem for Panda. She wasn't comfortable with people rushing up to her. She has now become very accepting of all the attention, but I still want her to have a way of telling a blind handler when she's got people surrounding her. She does this by backing up away from them and pressing closer to me. I can then decide if this is an appropriate time for petting. That will allow Ann to control access to Panda so she is not distracted from her work.

 

In December I expanded even further on the "leave it" game. "Leave it" is such a useful exercise. Dog trainers understand the necessity for it. There are all kinds of things dog owners want their dogs to leave alone, everything from their bedroom slippers to the disgusting pile of better-not-to-know-what-it-is that their dog wants to roll in. Horse people tend not to think about "leave it" games, but it is just as useful for our charges. Think of the lead ropes, reins, blankets hanging by the grooming area, hay bales, brush boxes, etc. we would like our dear, charming, ever-so-mouthy young charges to leave alone. Think of the tempting spring grass we would like them to ignore. "Leave it" is a most important lesson, and one we can learn from dog trainers.

 

Panda and I played many forms of the "leave it" game during our afternoon house sessions. After her snooze, I would tempt her with goodies. I began simply enough by holding a chunk of carrot out in my hand. If Panda reached for the carrot, the carrot disappeared. But if she hesitated at all, click, she got a treat.

 

The hesitation grew quickly into "you can't get me to touch that carrot." When I held it out in my hand, Panda would draw back in her pose. Click. Once she heard the click, she knew it was all right to take the carrot from my hand. But any grabbiness before the click simply made the carrot disappear.

 

Exercises like this are why experienced clicker trainers can say with confidence that hand-feeding does not create a mouthy, or pushy horse. Instead using food as a motivator can actually diminish the mouthiness. That's because the clicker let's us set up rules around the food. The horse learns that mugging pockets isn't the way to unlock the vending machine.

 

Remember the principle: "the horse doesn't know when it doesn't count, so it always has to count." If the handler is consistent and clear about the presentation of the click, adding food into the training builds good manners. If the handler is erratic in the timing of the click or gives food randomly as unearned treats, mugging may indeed be an issue. You do get what you reinforce. If you don't like your horse's behavior, look to your timing. Think about what you want your horse TO DO, and reinforce that. The more you focus on and reinforce the behavior you want, the more that's exactly what you will get.

 

Panda got quite good at the "leave it" game. We evolved from holding the carrot in my hand to setting it down on the floor, and eventually to scattering grain, carrots and cut up apples all over the floor. Panda would stand over the goodies posing for all she was worth. The pose makes a perfect counter-balance for the "leave it" game. It gave her something positive to do as an alternative to grabbing the goodies.

 

"Leave it" is still a work in progress. Panda is very good at drawing back from something once she understands that's the game we're playing, but she hasn't yet fully associated the verbal cue "Leave it" with the behavior, nor has she yet generalized it to real world situations. My goal behavior is to be able to walk through the high school cafeteria, feel her reach for something the kids have dropped, tell her to leave it and have her walk on. Since Ann works at a high school, that's something her guides need to be able to do. We're a long way from that target behavior, but that's all right, so are the guide dogs!. The principle is: "Never start with your goal. Put as many steps between you and your goal as you can." That's exactly what we're doing now.

 

Quarry

December finished off with two holiday excursions. The first was to Saratoga the Saturday before Christmas. I was again teaching over at Dolores'. After we finished the last lesson, four of us drove into Saratoga to go window shopping. Of course, we had Panda along for the training walk. Ann was with us, as well with her new guide dog, Quarry.

 

Quarry arrived Thanksgiving weekend, just a month after the dog she'd gotten earlier in the summer had to be returned to the school. That first dog had far too much prey drive to be a guide. She was constantly distracted by interesting scents, other dogs, children, sparkling lights and most especially cats. Anything in motion captivated her attention. Guide work was fun so long as there was nothing more interesting going on, but she was too easily distracted to be reliable or happy in the work. She has now been placed in a search and rescue program which is much more suited to her temperament.

 

Trying to make her work and finally coming to terms with having to send her back to the school was a frustrating and emotionally upsetting experience for Ann. When people ask why use a horse instead of a dog this is part of the reason. Ann's first guide, Bailey, was a wonderful dog who worked well right from the beginning. But as Ann has now learned first hand not all partnerships flow together with such ease.

 

As riders we can all appreciate that the beginning of a working partnership is sometimes not as smooth as we would like it to be. Eventually, we do indeed develop a rapport with our horse. They know their job, and we know what to expect from them, but it can take time to get to this point. A blind handler must develop trust in her dog to be able to move fluidly, especially in unfamiliar environments. Watching Ann with her new dog I have come to appreciate just how much time and energy must be invested in this process.

 

If horses can indeed be used as guides; if all the questions surrounding the practical day to day living with a horse in human environments are answered in the affirmative, the number of times a blind handler will have to go through this process will be greatly reduced. Never mind the emotional investment people make in their guides. The time and energy investment that's involved in replacing a guide is huge. That may be a major consideration in choosing a horse over a dog as a working partner.

 

Ann's new dog, Quarry, arrived Thanksgiving weekend. She is now in the adjustment period learning to work with him. Having Quarry in service buys us time for Panda. I'm under no pressure to put Panda into full work before she is ready. She is now just a year old, and both Ann and I still consider her to be very much in what the guide schools would call the "puppy raising" stage of her training. House breaking, socialization, basic manners, this has been the focus of her training to this point. Our plan is not to put her into full work until she is mature. We have plenty of time to explore all the questions we have surrounding the use of minis as guides, and to do a thorough job in her training. For us Panda's training is a research project. With Quarry in service for Ann, we have the time we need to investigate all our questions without rushing Panda.

 

Holiday Treats

So Quarry joined us for our holiday excursion to Saratoga. It was one of his first training runs, and a great opportunity for Ann to practice city work with him. For Panda it was only her second trip to a city. The first had been to downtown Boston in October.

 

Our holiday foursome was made up of Dolores Arste, Julie Varley, Ann, and myself. Our adventure began after we parked the car in a municipal parking lot. We had to go down a double set of stairs to reach the main shopping district. Panda, I am pleased to report, handled the stairs like a pro. She didn't even hesitate, even though the stairs were steep and in uneven lighting. I was thoroughly impressed by my little mountain goat!