Take a moment to think about your closest friends. What do they have in common? For me, there are three main things. First, my friends and I often do fun things together- whether it's a silly board game or tubing on the lake. Next, I find them endlessly interesting. We come from all sorts of backgrounds, so I'm always learning something new from them. Finally, and most importantly, my best friends are the ones who support me when I make mistakes, who encourage me to try harder, and who celebrate with me when I succeed.
My good friend Sara and I talking to Denise Fenzi.
Photo by another good friend, Robin Sallie.
While my answers are honest, they're also convenient, because many of the things I saw Denise do (or that she encouraged us to do) fit into these three categories. The way she trains lends itself well to developing great relationships with dogs, and I really think this is the part of the seminar that will change my training life.
Let's start by talking about play. Part of the reason Denise spent so much time talking about tugging is because it's a great way to play with your dog. But you can't take toys in the ring, so you need to find ways to play without them. On day two of the seminar, Denise had each of the working teams play without toys. This was easy for some, and harder for others, but Denise coached each pair on how to have fun together.
For example, she played a tag and chase game that many of the dogs loved- when the dog got into heel position, she would lightly touch his shoulder or chest, and then turn quickly and run away so the dog would chase her. She also encouraged the dogs to be physical; running and jumping is rewarding to most dogs. She did this by using things like the opposition reflex (where she pulled the dog back and then let him shoot away towards a jump or dumbbell) or hand touches in order to get the dog to jump in the air.
Play is a great way to reward the dog, but you need to make the work itself interesting, too. Simply drilling the exercises is boring, even if it is interspersed with play, and most dogs will droop under the weight of endless repetitions. Denise encouraged us to think outside the box- you don't need to train the exercises in a set order, nor do you need to do the full exercise every time you train. Instead, she recommended mixing things up each time so that the dog never knows what to expect. Sometimes you might recall the dog to sit in front of you, and other times you might have him run between your legs. Do signals backwards, or start from a down instead of a stand. Have the dog sit on recall instead of drop. The opportunities are endless.
While most people don't think about it this way, heeling really is just one long drill. Walking in straight lines? Not that exciting. Your job as the trainer is to mix it up, and Denise does this by walking erratically. She drifts left and right, makes frequent turns, and spins in circles. She does serpentines. She keeps it interesting by being unpredictable. And when she goes in the ring and walks in a straight line? Well, since that's so unusual compared to what she does in training, her dogs find it interesting too.
Finally, a good relationship will be supportive. One of the things that was so absolutely revolutionary to me was the way Denise used her voice in training. She is always talking to the dogs she's training. I am often rather quiet while training- that's what many of the leading clicker trainers recommend you do, after all. But one of Denise's cricitisms of clicker purists is that it uses silence as a “No Reward Marker.” In other words, clicker trainers tend to allow silence to let the dog know he hasn't got it quite right yet, and uses a marker- either a click or a verbal word like “Yes!”- to let him know when he's got it right.
This can backfire in the ring, where the handler must remain silent, even when the dog has it right. Denise's solution is both brilliant and obvious, and it makes me feel a bit stupid that I never thought of it myself. Simply put, she talks to the dog when he hasn't quite met her criteria yet, and then falls silent when he's got it right. After a few seconds (or longer, depending on the dog's level of training), she'll break the silence to reward him with praise and play. Thus, the silence begins to act as a secondary reinforcer by predicting exciting things. Genius.
But remember what I said about being supportive? It's important to note that she isn't scolding her dog or using a stern tone- she said that suppresses behavior instead of increasing it. Rather, she's encouraging the dog to try harder by being his own personal cheerleader. As I've begun to incorporate this into my heeling training with Maisy, I find myself saying things like, “Where's my puppy? I know she's around here somewhere! She should be right by my side!” in a very happy, upbeat tone. When she gets there, I'm quiet for a moment, and then BAM! I throw a verbal party.
In order to help the dog understand the difference between "you're almost there" and "you've got it," Denise said that you need to have two different levels of tone and energy when you're praising your dog. The “encouragement praise” is cheerful talk meant to keep the dog in the game, while the “party praise” should be over the top excitement. Both will be happy and positive, but it should be obvious to the casual bystander when the dog has gotten it right. In other words, Denise praises the dog for his efforts, and rewards the dog for success.
When the whole package is put together, what happens is that working becomes interesting and engaging because the handler is keeping things fresh and unpredictable. The dog remains motivated to keep trying because he's encouraged to do so. And when he gets it right, his reward is to celebrate with his trainer. Sometimes that means he is told how amazing he is, sometimes he gets a bit of food, and sometimes he gets to run and play. The common thread is that he always gets 100% of his person's attention, and the reward becomes more than just the tangible item he's given. It's about the interaction and relationship with his person.
Look, I know it sounds a little crazy. I don't entirely understand how it works myself- and I certainly can't explain it all in scientific terms- I just know that it does. Not only do Denise's accomplishments speak for themselves, but in a few short training sessions, I'm already seeing improvement in Maisy's enthusiasm and performance in both heeling and her dumbbell retrieve. Better yet- Maisy and I are having fun together.
What do you guys think? Do you talk during training, and how or when do you use silent? What troubles have you encountered? Do you think some of these ideas would help, or are you a bit skeptical? I'd love to hear your thoughts!
22 comments:
Well I can't know for sure but it sure seems like the new to us methods are helping. Magic scored 3 perfect scores doing rally outside, in the heat this weekend. I found myself changing my method in the rally ring to much more like the training we have been doing, a little nagging when out of place but trying, and using my self to get him up and going. At one point I even told him I was "gonna get him" and he perked right up. Loved it. Its working for us too.
It was fun to watch all the teams try it! Despite my failure and over powering my dog!
The silence=right and marking mistakes wasn't new to me for heeling, but I hadn't made those same conclusions for every other exercise! I feel kinda silly for being so inconsistent with my dogs!
the talking thing is interesting ... I hate HATE cheerleading .. so many agility people fall into the trap - but that's not what Denise is talking about at all ... I LOVE the use of silence and try hard to use it but I am fairly vocal by nature so with Brody sometimes fall into using sounds ... like trains or squeals - I don't even realize it but apparently it's motivating for him and entertaining as all get out for the audience ... I don''t have a lot of time to make noiseof any sort with Sally :)
I've got a bad problem with being quiet when I'm concentrating - but Lola likes it when I chatter! I really need to work on building up her enthusiasm for my silence, and my own ability to cheerlead for her when she does things right. :)
Dragon likes it when I talk to him and encourage him when he's a bit frustrated. I've also been making a consistent effort to quietly smile at him when he's doing something right, in order to build that as a secondary reinforcer. Of course I always follow it up with a primary reinforcer and praise. So I think that, without realizing it, I was already implementing some of Denise's program! That's probably thanks to reading her posts on the clickcompobed list, actually. She really is brilliant! :)
Elo is my easiest dog to train and motivate, because he finds just being around me, being touched, and playing with me motivating. He likes chase games and roughhousing. Jun is harder to play with without a toy, but I've been experimenting and trying to find games she likes. She likes hand targeting and jumping up for a target. Last night we were playing "chase me to a hand target" and she really started to get into it when I pulled it away from her at the last second a few times before letting her get it.
Having a deaf dog, I do talk to her, but not a whole lot. I've kind of carried that over to Elo, but for the past few days I've been trying to talk as encouragement when he's not quite there and be silent when he gets it right. Too soon to drawn any conclusions, but he seems to like it. I wonder if he would become dependent on the cheerleading to keep trying though?
Hi! I'm a long time reader but I don't think I have ever posted before. I really enjoy your blog. I think the idea of silence being a secondary reinforcer is brilliant and I love the novelty of Denise's approach. But thinking of the quadrants, can one really argue that going silent is adding something? It seems to me that from our human perspective we can "define" the talk as encouraging and supportive, but if the end of the talk signals a correct behavior, isn't this a negative reinforcement protocol? (Followed by positive reinforcement.) I'd love to know what you think.
Oh my gosh, E, I LOVE YOU. A fellow science geek!
I definitely have some thoughts on this, but I'm just leaving for work now, so I'll reply later this evening.
E, i would agree with you that the "supportive" talking is a correction and is R-. The dogs know the difference between your true happiness and your happily correcting them. I also think that is why training this way can be very effective! If trying to train as positively as possible it is really hard to give the dog feedback that isn't abusive in any way and R- training certainly is effective. The verbal feedback, done in a way that doesn't stress the dog is a great way to take advantage of R- training :)
Okay, E, here are my thoughts:
First, I think there's a problem of terminology. When I describe Denise as a "positive trainer," I'm not talking about where she falls in the quadrants, but rather, the fact that she avoids pain and fear in training. I think this is a habit a lot of positive trainers have.
Second, I'm not sure Denise would care much which quadrant she's technically in- there's a place for science, of course (I love it, personally), but my sense is that she's more of a "get the job done" person. I think those things can overlap, but don't have to. I also think that getting stuck in a certain quadrant/box can hinder training. (It's happened to me.)
Third, and finally to your point:
If the trainer has conditioned silence well, and the dog has come to expect something good when he hears silence, then what the dog wants is for the talking to be turned off and for you to give him silence. In that case, I think there are two ways of looking at it scientifically:
1. You are removing something (talking) to increase the dog's behavior (heeling, for example), thus R-, or
2. You are adding something (silence) to increase the behavior, thus R+.
Since silence is typically defined by what it isn't (noise), the second seems more spurious. So, yeah... it's probably technically a negative reinforcement protocol, which is really interesting since R- is considered "bad" since you have to start something the dog wants to go away. I actually think it's pretty clever to be able to use praise in that way, but I do think it requires a really solid relationship. As Laura points out, it will only work if the dog can differentiate between "bad praise" and "good praise," which means you need to be able to genuinely praise the dog.
Does that make sense? I'm afraid I'm getting lost in my argument. :)
wow; that was a complicated set of comments I just read. And yes Crystal, I just want to get the job done:).
Here's how I percieve it from the dog's point of view: When I'm "thinking" about stopping work for a toy or a piece of food or a play session, I stop talking and smooth out the work. The dog has learned from experience that this silence tends to preceed a very good reward, so they perk right up in the face of silence.
But....I also use two distict forms of praise with my dogs; the "you're trying hard please stay in the game" praise, and the "you got it right and are brilliant and I might go ahead and give you something else you want" praise. So....it's not like I'm always silent before I reward, but I am NEVER silent if the dog is in the process of failing miserably, because that is sort of mean to the dog. It leaves them hanging and alone when they need you, and as a result silence is quickly poisoned in the eyes of the dog. That has very poor ring carryover. Here's a video I took today of a young dog working through gloves. Note the tones of voice I use - I'm always engaged and praising but some praise is "you got it right" and some is "you're a good dog keep trying". and note that on one rep when she is correct, I'm silent until the last moment and then I reward directly with the toy. In that case, I must have had some reason to believe that getting it right was becoming a habit, so I no longer felt we had to have a party for each correct rep. Anyway, here's the link to the video: http://www.youtube.com/watch?v=_xM-yIHo8D0
Apparently posting links gets you marked as spam...
Thanks for the comment, Denise. :) Going geeky again, I think it's interesting how you have praise the dog wants to turn off and praise the dog wants to turn on. When I first saw your videos, I really couldn't understand how in the heck your method worked. I still don't, really- but I like it, and it's working for us, so I'm trying not to question the science too much and just roll with it instead.
I don't' think the dog wants to "turn off" the encouraging praise. It simply is a much lesser form of reward than what silence leads to, but both are appreciated.
Think of it like this. If someone were teaching you flashcards of multiplication, and you got candy when you got five in a row right, wouldn't you prefer is they said, "wow; great try but not quite right" when you made a mistake rather than staying quiet or looking glum? then, when you got to three or four in a row, wouldn't in be sort of normal for tension to build as you got near the fifth correct one (which both of you knows is the winner) and silence to ensue? think of any contest with a big winner at the end. As the tension grows and the contest gets closer to a winner, eveyone gets quiet and intense.
Beleive me, I know all about this because during the school year I am tortured with practicing spelling tests with my son. All he gets at the end when they are correct is the end of the torture, but still....I encourage a lot more when he's struggling so he'll keep trying. When he's getting it right I'm more quiet. If I get excited over every correct word, then it loses it's meaning - becomes background noise. Instead I'll be quiet if he starts getting on a roll and occasionally say something quiet like "you're doing it".
Denise, you're wrecking my theory on the science behind your method...
I guess my question now is: is encouragement rewarding? It's certainly interesting and engaging. Is it rewarding the effort? Or just spurring the dog to keep trying? Or both?
I suppose as long as it works, it doesn't matter, but the geek in me wants to know!
crystal, I will happily continue these conversations until you feel you get it. Everything I do is logical and can be explained. Sometimes I just need someone to force me to explain it or to point out what isn't understood. But the logic is always there.
I think any time I'm with my dogs and paying attention to them, then it's rewarding. People too - I should point out that my son could do his spelling words alone or on the computer, but he wants the interaction so I get sucked into that job.
so...is encouragement rewarding? I'd say yes, it is rewarding, because it is focused attention. the alternative is sitting in the house doing not much of anything.
I would be willing to bet that HOW rewarding varies wildly by dog. I know that the dogs I have taken the time to build a strong interpersonal relationship with (not the food/toy part; just the "me" part that emphasizes company and pleasant interactions) get more reward from being encouraged than a dog I dont have a relationship with.
for example, I would not try and take someone else's dog and just use a ton of encouragement without the food/toys, because why should the dog care about me or my praise/encouragement? But any dog that I see over time that I make an effort to have a personal relationship with....those dogs will work for the time to be spent with me. But....it is a rare dog that takes praise over food and toys. Obviously the best option is the package deal, where you tell your dog he's brilliant while feeding and playing:)
You're so awesome, Denise.
I feel like I have a decent handle on what you do, I just really need to get out there and practice doing it. I think that the more I do it, the better I'll understand it. (This is not to say that I think I have anything near the brilliance you show in training. But I'm going to practice and get close!)
What I don't understand is how it works scientifically. If the entire interaction is enjoyable, it doesn't seem like it's negative reinforcement, like we were speculating... at times it is, I think- like when you use the more "harassing" tones- but for the most part, it just feels like some things are MORE rewarding.
Of course, there is a difference between "rewarding" and "reinforcing" (enjoyable vs. increases behavior), and I guess what I'm getting from you- both in terms of what I saw, and what you describe with your son's spelling- is that what reinforces the behavior is the relationship and the interaction. In a way, it seems like getting it right (even if that behavior seems silly to the boy/dog), is reinforcing because YOU want the dog to get it right... because the dog wants to make you happy, because that makes him happy, which is so incredibly anthropomorphic and unscientific but makes SO MUCH SENSE.
Sometimes I think science underestimates how much we can bond with our animals, and how important that relationship is to BOTH parties.
Keep in mind that if I ever used negative reinforcement that truly disturbed the dog, they would leave the training situation because I don't "make" them stay. This is the part about choice that is highly relevant. Because the dogs can choose to leave whenever they want, as can my student's dogs, the TRAINERS are highly motivated to figure out how to make the entire session fun. Or at least fun enough that when some mild pressure/harassment or endurance expectation is placed on the dog that they choose to stay in the game.
Your comment about science is my long standing complaint about much of what I see, read and hear about science based training as it relates to dogs. There is a basic belief that anyone with good technical skills can teach any animal/dog almost anything. that might be true with chickens and pockets full of food, but with a dog you have a few distinct differences. 1) They meet us more than half way, so even with less than stellar training they'll often compensate for the owners. Sometimes stunningly so, which unfortunately leads some trainers to believe that they actually know what they are doing when they do not, and also lends credence to their belief that the dog is blowing them off when the dog does NOT compensate for them, 2) dogs naturally care about the underlying relationship with a person who spends time with them. So while anyone can work any of my dogs if they have a fist full of food or a toy in their pocket, I can guarantee you that if you don't have those things readily available then my dogs will NOT work for you, because they really don't care about you or how you're feeling about them. and 3), many dogs truly seem to enjoy learning and work for the sake of the mental activity, something I've never seen addressed in a scientific fashion. So regardless of whether or not there is an external motivator being applied, I think some dogs simply enjoy challenge for the sake of challenge. I think some (most?) dogs can have a sense of pride or accomplishment, and we can foster that in them.
Anthropomorphic? You bet. But does it explain behavior like why my dogs prefer negative attention (harassment in training) to no attention? Yeah, it does.
Taking care of Juno has truly brought this home to me. I take care of her four days a week. I've always taken good care of her. I am nice to her, give her treats and snacks and train her every day. But I didn't take that extra time to pull her out by herself, and let her know how special she is. And as a result, she learned to work strictly for toys and food - she might as well have belonged to someone else. And when I saw this I changed our relationship. I started making a point of spending special time with her, just petting to her and talking to her and letting her know that she was special. I would push my other dogs away so she could come in for personal attention. And what a difference this has made in our work. She is finally starting to work FOR ME. it's a long road and we have a long way to go, but she taught me without question that you cannot get maximmum performance out of a dog if the dog doesn't feel he or she is a very special and important team member.
That reminds me a lot of what you said at the seminar, Denise: For a true training challenge, SHOW.
I don't think there's anything wrong with studying the science- I really enjoy learning theory- but wasn't it the Brelands who found that sometimes it falls apart despite your best efforts? I'm going to have to read "The Misbehavior of Organisms" again, but I believe the gist of it is that sometimes instinctual behaviors override conditioned ones.
How does this apply to obedience? Honestly, I have no idea. But I think thousands of years of evolution and purposeful breeding have primed our dogs to want to work with and even- gasp- please the people they are in relationship with.
And your point about dogs wanting to learn for the sake of learning? I don't see why that couldn't be true. Humans enjoy learning for fun (heaven knows that's why I read and go to as many seminars as I do), and our brain structures are very similar. Just because science can't prove something now, doesn't mean it won't in the future...
I'm still going to pick apart the science- it's too fun not to- but somehow, being able to let go of all that DURING TRAINING is so much better. We're having a lot more fun these days, anyway. I am so very glad we attended your seminar.
Something I just saw on Facebook mentioned the famous behaviorist BF Skinner, which reminded me of much of what worries me about science as it relates to dogs and dog training for performance. Remember that in his time, BF Skinner was the GOD of behavior and if he said it, then it was true.
BF Skinner advocated putting babies on strict feeding schedules and not coddling them when they cried (to avoid spoiling). What came of that advice was a condition called "Failure to thrive". basically, that means that if you dont' love your baby and pick him up a lot and make him feel loved, then he'll die even with food and warmth and all of his physical needs being met. The moms who followed these schedules were doing what science told them, even though their gut instinct told them to hold their babies and comfort them when they cried.
Another researcher named Harlow demostrated the same basic principles with baby monkeys....if they aren't loved by a mommy monkey, they end up very messed up monkeys indeed.
So....is it such a stretch to believe that our dogs need love and that feeling of being special to give us their best in training? Are reward schedules and quietly handing out tidbits the way to get the best behaviors that will hold up long term under stress and when the external reinforcers are no longer there? Maybe cookies and toys can give us behaviors, but do we risk losing the essence of relationship if we don't' make the absolute value of work deeper than the quantifiable elements.
In 2004, I was working and living in a group home with four people with pretty severe disabilities. When you live with and care for people day in and day out, you grow close to them.
One day in early September, one of the women, whom I'll call P, was sick, and so the staff working took her in to the doctor. They diagnosed either bronchitis or pneumonia- I genuinely don't remember anymore- and sent her home on antibiotics.
The next day, I was working, and noticed that she wasn't well. Unfortunately, P was nonverbal and had no functional communication, so I couldn't ask her how she felt. So, I called the on-call nurse. "Crystal," she said, "of course she isn't well. That's why she's on antibiotics." I hung up. I called again several hours later, and then again after that, announcing that I was going to take her to urgent care.
When we arrived, the doctor told me the same thing. The antibiotics needed time to kick in. And anyway, he said, all her vitals are fine. They're normal. Go home, let her rest.
I didn't take her home. I KNEW something was wrong. Seriously, seriously wrong. I took her to the ER. It didn't take us long to get there, but by the time we did, her oxygen saturation in her blood had dropped quite low- something I could have never discovered if I'd taken her home. She was admitted that night, and remained in the hospital for three weeks.
Scientific wisdom said there was no need for care. Scientific wisdom said I was overreacting. But my gut knew better. I knew that something was seriously, seriously wrong. Later, I was asked how I knew. I didn't know. I still don't. But I'm sure it had to do with our relationship.
Science is wonderful. It's fascinating. It revolutionizes the way we do things. But P's life (and ultimately, death), changed my life. Relationships are what matter.
I agree with you Crystal.
I also believe in science and find it quite interesting; I seriously considered the "academic route" after college so that I could be a researcher. And science has done more to "humanize" and civilize dog training methods than anything else I know of.
Now I'd like a harder look taken at the wierd stuff. dogs that work in spite of terrible abuse....and look happpy doing it. How and why does that happen???!!!!
I have wondered that too! And why do some clicker dogs look so miserable?
Theory: Relationships affect the final product, regardless of training methods.
Now we just need to start a study!
Post a Comment