Motivation: Extrinsic vs Intrinsic

[mbc1963]

I guess I just don't see the big divide that some people see.

 

I never used a clicker before, because I just couldn't see having to keep yet another thing in my hands when I was already carrying a leash and a bag of poop. (How do people manage that?) With this new girl Cricket, I decided to try clicker training, but I just use my mouth to make the click: that little "git going" horse sound people make. So, I "loaded" the click for a few days with lots of treats, and now I use it to mark good behaviors. I still use treats to teach a specific behavior: Cricket learned "paw" in about three sessions a couple days ago. For this specific purpose (letting the dog know at the exact moment she starts catching on to the behavior I want), I find the loaded clicker thing to be miraculous. It's so direct, such purposeful communicating with the dog.

 

Meanwhile, though, I take her off leash a lot. She's been with me nine months. In the last couple months I've realized she's highly reliable to come and stay right by my side if we're passing a human in the woods. I just say "over here," and hold my hand by my side, and she comes over and walks by my hand. She'll likewise come to me immediately when I need to leash her, and she'll wait when I say "wait." I didn't train any of this with the clicker or even with deliberateness; it's just body language and casual repeated use of English in everyday situations. In this sense, she's doing what I ask "because she wants to." But she wants to because we have a good relationship and she trusts me.

 

Using treats makes things fun and exciting and makes the dog happy to have a session. I don't think that excludes the building of an intrinsic desire to do as the owner expects.

[D'Elle]

Even though I had - absolutely superfluously - lavished treats on him (one treat for every correct response, and most incorrect responses!) for years, when the time came when I needed to pare down the use of treats, I was able to do so.

 

How can that be?

 

Because when the time came to reduce reinforcers, I approached it as teaching him something new.

 

I wasn't thinking in terms of "getting rid of the food". I think a lot of new trainers get into trouble when they think in those terms because they end up trying to rush, or they raise criteria way too high. Instead, I was thinking in terms of "new behavior".

 

Personally, I err on the side of using too much food (or toys) for too long because I know that I can - very effectively - reduce the use of such reinforcers when the appropriate time comes to do so.

 

And if their use is superfluous for a time . . . so what? As long as we get the desired results in the end, it's all good.

 

Granted, having experience with this now, I handle it smarter. But there are still times when I need to make a judgment call in this regard. When in doubt, I provide the extra motivation.

Treats can be a distraction from the training, but only if they are constantly in the trainer's hand. If they are kept out of sight and only come out for the reward, they are not a distraction.

In Freestyle work, especially competition, as Root Beer says above, it is imperative to get the dog to work without treats. An excellent Freestyle trainer who cane to give a seminar here that I and my freestyle friends attended said that fading the treats doesn't work as well if you do it in the way most people would think. Many people think the way to get the dog to work without food reward is to ask the dog for two behaviors before treating, then for three, and so on. It turns out, according to this trainer and backed up by the experience of all of us who have tried both methods, that the exact opposite is true. Treat the dog like crazy in training, and even just before entering the ring. Don't run through the entire routine at any time except in performance. Then do the routine without treats and treat heavily once you are done. The trainer said that it is like putting something "into the bank"....you bank up on the treats, and "make a withdrawal" when it is performance time. I did not think it would work, but it works like a charm, and is the method that all of us use now if we want to perform with out treats.

Of course, this is only about freestyle. When it comes to many other things, like agility and especially sheep, the work itself is the reward. And for some dogs freestyle is the reward as well. It just depends on the individual. Like Root Beer, I will always err on the side of "too much" treating when I am training, unless I am training a dog who would rather work for playtime, or who is not food motivated.

 

Another thing that is somewhat counter-intuitive is that intermittent reinforcement often works as well or better than consistent reinforcement once the desired behavior is learned. The dog doesn't know whether or not he will get a treat, and so is eager to do what I want in order to find out. I am careful how I use this, though. Close observation of the individual dog is important, since of course one size doesn't fit all when it comes to dog training.

 

But I also agree heartily with those who are saying that respect, trust, consistency and the relationship are the paramount building blocks in training of any kind. Without those things as a foundation, you won't succeed. Suzanne Clothier ("Bones Would Rain From The Sky") is the founder of "Relationship Centered Training". She said in a workshop I took from her that she could tell just watching people get out of their cars with their dogs what kind of relationship they had and what kind of trainers they were. I think I can do the same to some extent when I see new students come into a freestyle class. The body language of the person and the dog tell everything.

 

Like Cass C, there are some things that, although well learned, I will always spend time reinforcing, and the main one is recall because it is the most important for the dogs' safety. My dogs have very good recalls, but periodically I will reinforce it by passing out very yummy treats when they come.

[Alfreda]

A lot of that has practically nothing to do with training theory and or practice. If your body language makes sense to the dog it doesn't matter what silly ideas you maintain.

 

I can't tell whose dog is going to go well when handler and dog walk to the post but I can tell who will do very badly. And I'm not even a dog.

 

Donald McCaig

Haha! love this!

 

Which reminds me of an aspect of BODY Language that's also interesting in this context of "intrinsic vs extrinsic": it goes both ways. Body language reflects an inner (possibly unconscious) state, but also, it can apparently be used intentionally on the "outside" to alter an inner state.

 

In the TED talk below, Amy Cuddy presents her research on assertive body language. People who might lack (inner) confidence can apparently build that capacity by intentionally adopting more confident body postures. Doing so may start out as a mere "act,” but gradually the outward posture begins to change the inner state. This is something that can be measured internally (hormones), as well as outwardly, perception of others. (See TED talk by Amy Cuddy)

 

https://www.ted.com/talks/amy_cuddy_your_body_language_shapes_who_you_are?language=en

[Donald McCaig]

Dear Doggers,

Writing about effective teachers, Bill said: "My job was to help teachers improve, but on the occasions when I could predict poor performance with a single glance, it was a difficut task indeed."

 

We are much better at this sort of thing than we credit. From my profession: Some years back I was sitting in my literary agent's office on a Monday morning when all the unsolicited manuscripts came in. He tossed me half a dozen envelopes each of which contained three hundred or so pages of someone's literary hopes and dreams. Ten minutes later I said, "Nope."

 

 

 

Donald McCaig

[Maralynn]

Sure treats can used to reward behaviors and are great at instant gratification, especially for pups whos attention span is limited. But along with treats should be praise then you can easily phase treats out quickly leaving reward as praise.

 

I think there is a huge difference in the dogs mindset if they are working for approval/ to please you or if working for food and self gratification. If for food they can tell quickly if you have any and some will simply decide it is not worth it.

 

If using food correctly in training, dogs just don't decide that it's not worth it IME.

 

My dogs refuse food if they're over threshold, hurt, scared or highly confused. In regular training? No way! They work for kibble. In higher distraction areas (vet's office) I may use cheese. But I don't rely on the food to train, I use use the food to add value.

 

So perhaps this is an example of the relationship being at the foundation of training instead of the food that is also being used.

[WickAndArtoo]

My dogs like praise, thank gosh since I am very clumsy with treats and clickers lol it took me a long time and lots of treats (and money) to realize wick actually likes praise better (it's hard to tell because he basically ignores you...but yet he tries harder when treats aren't involved). Since I've learned this our bond has strengthened ENORMOUSLY and his learning has been much faster!

 

That being said treats have their place and many uses that can't be done with just praise as easily. Luring & treat play and many other situations can have majorly fun outcomes!

[Root Beer]

If using food correctly in training, dogs just don't decide that it's not worth it IME.

 

Yeah. I can honestly say, I have never had that happen. And I never stop training my dogs. I have video of Speedy working on a pivot platform the week before he died. He was 12+ at the time. Dean is 10 and has yet to become uninterested in food in training.

 

When food is used well in training, I find that I can toggle between having treats and not having treats without much difficulty, but I have never had a dog actually lose interest in food altogether.

 

In my experience, use of food is part of a training language that I build with my dogs. The food can communicate many things - some of them quite nuanced. Food, toys, use of environmental reinforcers, praise, verbal encouragement, body language, targets, verbal and physical cues, training props, the clicker, eyes, etc . . . all of it forms one big "package" that I can use to communicate with my dog in a training context.

 

I have yet to find that my dogs get to a point where they consider any of it to be "not worth it". In fact, by the way they all vie to be the one who gets to train, I'd say they all consider it to be very much "worth it".

[Root Beer]

 

Because when the time came to reduce reinforcers, I approached it as teaching him something new.

I wasn't thinking in terms of "getting rid of the food". I think a lot of new trainers get into trouble when they think in those terms because they end up trying to rush, or they raise criteria way too high. Instead, I was thinking in terms of "new behavior". - Root Beer, Could you give me a practical example of how this worked?

thanks

Bill

 

Sure.

 

Let's say my dog can move with me, with good attention, for two minutes. Now I need to help the dog transition to the point where he can do so with me having no food on me.

 

I am going to approach "moving with me for two minutes without food" as a brand new behavior. So, I am not simply going to go on the floor with my dog, without food, and expect two minutes of movement with attention.

 

I will start by breaking our work into smaller pieces. Say 15 seconds. I will start with food on me, but go to reinforce from a remote location. So, the dog starts to learn that reinforcement might come from somewhere other than myself.

 

Once the dog is successful with that, we will give just that much a go with me having no food, and going to reinforce from that remote location. If the dog loses focus, I will go back to working with food on me, but reduce the time moving even more.

 

After a successful repetition with me having no food, I will get food and do a repetition with it. I am starting to help the dog see that whether I have food on me or not, it's the same thing.

 

We will work at that stage, until it really does not matter. I will also change up the remote location from which I reinforce. I might move that into a different room, even. If the dog is toy motivated, I might also use a toy instead of treats as the remote reinforcer. It just depends.

 

Once that is easy, we will start the whole thing over, but expand to 25 seconds. I am making these times up. I might increase a lot more gradually if the dog needs that. But sometimes these kind of time jumps are fine.

 

Rinse, repeat steps above.

 

As we continue this, gradually building to 2 minutes, I won't always make it more difficult. Some days we will reduce the time.

 

I might also mix up what we are actually doing. We might work simple movement on some days, but practice simple Freestyle moves (like twirls) on other days. Same principle.

 

Does that make sense?

 

What I am not going to do is just stop working without food and expect the dog to understand that we are doing the same thing that we did with food. I am going to use a process like this to help bridge the dog's understanding.

[ourwully]

Root Beer - Thankyou very much for the detailed response. It makes perfect sense and I will try to adapt it in some of the things I am doing.

cheers

Bill

[rushdoggie]

Root Beer - Thankyou very much for the detailed response. It makes perfect sense and I will try to adapt it in some of the things I am doing.

cheers

Bill

 

Bill, there's a new book out called Beyond The BackYard by Denise Fenzi that is full of very clear exercises on how to do exactly that. If you like having directions and a guide, might be a good read for you.

[Root Beer]

I second that recommendation!! It's an EXCELLENT book!!!!

[ourwully]

Thanks again. I will try and get a copy.

cheers

Bill

[ourwully]

I have just ordered the book on Amazon. It had a five star review but recommendations from you two are more than enough. Thanks again.

Bill

[ourwully]

How many people take the time to do both; a cookie, followed by praise, with equal energy?

 

It is funny how you can fool yourself. When I first read the Fenzi article I said to myself "Of course I am giving equal praise" but this morning I called Juno over on our walk and when she came over I automatically gave her a treat. I hardly gave her a second glance other than to give her the treat. I am not sure why I noticed this today but it was a real eye opener. I really thought that I was giving equal praise but in actual fact I was just handing out treats. Juno is very important to me and I spend a lot of time with her so I assumed that I was giving her my full attention during recalls. I had actually fooled myself into thinking I was doing something I wasn't. Anyway, I am very appreciative of the Fenzi article now that I have caught myself so now I will have to monitor my behaviour more carefully. My guess is that I am not the only one who has fallen into this trap.

Bill

[Root Beer]

You're right. I fall into that, too. If we are doing a training session and I want to work on, say, five different behaviors, I don't even think to stop and actually give praise all that often. I'm so focused on what we are doing, and reinforcing with food - usually.

 

It is a good reminder!!

[Pam Wolf]

There is a way to turn a behavior into an intrinsic reward. Via shaping a behavior, it is possible to turn a behavior into a reward for the dog, thus enabling you to use less food/toys during the training.

 

When training my Lassie for obedience, I often used the weave poles or jump chute as a reward for doing something she had had trouble with. After doing the difficult behavior I would simply tell her "go weave" or point to a series of jumps and send her-she would run across the yard to do the obstacles, come back looking quite proud of herself.

 

As a long time food trainer, I learned early on the need to fade treats and put them on a variable reward schedule only for better and better performance of a behavior. Also it is handy to have a variety of rewards to get better behaviours from the dog.

 

What Root Beer describes is shaping and it works

[CptJack]

There is a way to turn a behavior into an intrinsic reward. Via shaping a behavior, it is possible to turn a behavior into a reward for the dog, thus enabling you to use less food/toys during the training.

 

When training my Lassie for obedience, I often used the weave poles or jump chute as a reward for doing something she had had trouble with. After doing the difficult behavior I would simply tell her "go weave" or point to a series of jumps and send her-she would run across the yard to do the obstacles, come back looking quite proud of herself.

 

As a long time food trainer, I learned early on the need to fade treats and put them on a variable reward schedule only for better and better performance of a behavior. Also it is handy to have a variety of rewards to get better behaviours from the dog.

 

What Root Beer describes is shaping and it works

 

Yes. This.

 

My dog's 'reward' sometimes is spin. Because it's got a loooot of reward history with food and she thinks it's fun. That said, there's also VALUE TRANSFER from external reward to activity that I don't think I've seen mentioned, here. No, you don't have to keep using the food/toy forever, but you can sure as heck use the reward history/history of rewarding to build value into the task/activity/command performance.

 

 

[ourwully]

rushdoggie/root beer - I just received "Beyond the Back Yard" by Denise Fenzi. I haven't started the exercises yet but I have read the first two parts. I am going to start the first exercise tonight and work my way through the book over the next month or so. I will give a report on the book later as I see how things are going but I can tell right from the start that it will suit me. It is very clearly written and I like the way it works towards training without treats.

thanks

Bill