Dog Training Basics: Getting the Behavior

Tan puppy with black muzzle is lying on a navy blue bath mat and looks seriousI think one of the hardest steps for people who cross over to positive reinforcement based training is learning how to get a dog to start performing a behavior.

If we have experience with mild force-based methods, such as verbally telling the dog to sit, then pushing his butt down, or even if we have done lots of luring, it’s hard to imagine how to explain to a dog what we want them to do without taking one of those actions. It’s even harder to believe that he will do it repeatedly without a lot of chatter on our parts.

Getting a dog to do something over and over without saying a word, which can be quite easy once you know how, is very hard to explain to someone who has never done it or seen it. I have seen countless questions about it on Yahoo groups and FaceBook, and even after it is described to questioners, the videos that they subsequently post usually show them struggling to understand the concept. They speak verbal cues long before the dog understands the connection of the cue and the behavior, then wait impatiently for the stunned dog to do something.  Then they go through all sorts of gyrations to get the dog to do stuff. They also frequently startle their dog, are too loud, use pressure-ful body language, and do other things that cause the dog to not have a good time.

I feel OK pointing this out because I have done all of the above. Some of them you will see in the videos below.

I recently ran across this series of videos I posted when Clara was a puppy that document the first 5 sessions of teaching her “down” and thought they might be helpful to people who are not familiar with the process of getting behavior.

It’s not exemplary training, but that’s one of the points of my blog. The process is robust, and dogs are very forgiving. I do show my mistakes, and hopefully how I solve the problems I create. One of the four videos shows a miserable failure of a training session, but it’s a good lesson and also pretty funny. Clara is a little stressy in some other sessions. But I do a few things right, too, and they might be helpful as well.

The two things that I believe the videos demonstrate fairly well are a way to “get the behavior,” i.e. get a dog started on and repeating a behavior, and how to start adding a verbal cue.

Getting Started

My go-to method for getting and training basic behaviors from a dog I live with is capturing. It helps built great habits for me, since it makes me watch for my dogs to do something I like, and it is generally fun and very low stress for the dogs, too.

When you set up a training session based on capturing a behavior, antecedents are especially important. There are three types of antecedents: cues, setting factors, and motivating operations.  Cues signal that a certain behavior is likely to be reinforced. Setting factors make the behavior easier to do. Motivating operations provide an impetus for the animal to do it. You can modify the environment and situation through these methods to make the behavior you want easy for the dog to do and hence more likely.

In the case of lying down, I had already reinforced baby Clara for lying down on mats I have scattered around my house as “parking places” for dogs. I have entirely hardwood, tile, and cement floors, so the comfort of a mat on the floor is already an antecedent that sets the scene for the dog to sit or lie down. The mats are both cues for lying down, because of the reinforcement history tied to them, and also setting factors. Lying down on them is more comfortable than on the surrounding floor. I didn’t use a motivating operation, but an obvious one would have been to do the sessions after Clara had exercised and was a little tired. To me that’s a little bit of overkill; I’d rather have an alert puppy and use other ways to make lying down likely. So I used a mat to “jump start” Clara to start offering downs.

If you don’t use mats in your house, you could take your pup to the place where it is most likely that she will lie down. Perhaps her bed (or your bed!). Or you can do the “bathroom” trick. Go into the bathroom with the pup, have a bathmat down, and wait around and be boring until she lies down out of sheer boredom.

A small black and white rat terrier is almost lying down, but her chest and elbows are not touching the floor

Cricket demos the dreaded “bridge” instead of a down. Her chest and elbows are elevated.

(Capturing a down is not ideal for little dogs who hate to lie down. It can be because of the surface, or because they don’t feel safe in that position, or a combination of reasons. You probably already know if you have one of those dogs. Some of them will do the “bridge” approximation like Cricket in the photo.) Capturing could work if you can get them somewhere where they are very comfortable and relaxed, but you might want to use a different force free method instead, like shaping or luring a few repetitions. )

Applying the Cue

This is the other thing that is very hard to comprehend when you come from a traditional training background. We think we need to tell the dog what to do, then make them or get them to do it. To truly cross over, we need to let go of that idea completely and do things the other way around.

One way to help yourself get over the hump is to think of the cue as a label. (This is not technically correct, but it is a stepping stone away from the idea of a “command.”) As Sue Ailsby puts it, our message to the dog and in our heads can be, “That thing you are doing–we are going to call it ‘Down.'”

Again, almost everybody wants to talk first. Most of us have an unconscious belief that the dog understands the trainer’s native human language. And they are so good at reading us that they appear to understand much more language than they actually do.

One way that helped me finally get it that the “meaning” of cues is for us and us alone, was to see the silly cues that people sometimes give to behavior. You don’t have to call a sit “Sit.” You could call it “triangle” or “rocket” or “potaperm” and it would work equally well as a cue for the dog, if you taught it properly. And in fact, doing that–training a trick and giving it a nonsense cue–is an excellent way to see things a little more from the dog’s perspective. We operate from the meaning of the word. They have to go by pure sound recognition and memory.

The usual method of beginning to teach a cue is to get the dog doing the behavior repeatedly and consistently, then start saying the cue just as, or a hair before they perform the behavior. You’ll see that below in Session 4 in the video. My timing is not perfect.  I do it just a tiny bit early the first time, but luckily Clara, though startled, downs anyway.

The Training Sessions

What follows are five training sessions combined onto four videos. I’ve put a little explanation before each one.

Session 1 In order go get Clara “thinking about downs” I start with her on a mat. Then as quickly as I can I take away the mat but work in the same place. I look at her expectantly and she tries it! After she is doing it repeatedly I move us to a different place.

Link to the Session 1 video for email subscribers.

Session 2 First I do some review repetitions in the original location. Then I make a series of mistakes. I take Clara out in the back yard, and it is probably too soon for that more distracting environment. I make it worse by switching to a lower value treat (thinking that it will be easier to see in the grass) that she doesn’t even like. Oh oh! The session goes south. Then I even try to continue after she gets worried about a noise.

Link to the Session 2 video for email subscribers.

Sessions 3 & 4 (on same video)

In Session 3 we work inside again and I have her complete attention. She starts offering me a little duration, so I alternate with feeding her in position and tossing the treat to reset. Note that she is always free to get up after I click, but she sometimes chooses not to.

In Session 4 I start adding the cue as she is in the process of going down. As I mention in the video, she doesn’t “know” the cue yet; it’s just a noise that it happening in the training session. In retrospect I am not pleased with her demeanor in this session; she looks stressy and isn’t having much fun. Insensitive trainer!

In the second part of Session 4 we review “Sit.” If you do one behavior exclusively, especially with a pup, it tends to overshadow everything else. She did know the verbal cue for Sit pretty well, so we did a few successful reps. She was a little perkier, but I still went on at least one rep after she was saying it was time to quit.

Link to the Sessions 3-4 video for email subscribers.

Session 5

In this final session she is offering downs fast and furious so I add the cue again. I think it’s interesting that she actually startles the first time I give the cue (plus I was a tiny bit too early). She looked up and said, “Huh?” Then tried lying down. This session went on a tiny bit too long as well.

I’m always learning as a trainer and I think I am better now at gauging my dogs’ happiness with our activities and judging how long we should go on.

Link to the Session 5 video for email subscribers.

All Grown Up

I’m very grateful that even though Clara has such a short coat, she does not mind lying down on hard surfaces, as many smaller dogs do. Ahem–Cricket–ahem–Zani. It is a strong default behavior for her and we use it all the time.

Clara offers a down with eye contact on a field trip

Clara offers a down with eye contact on a field trip

 A Note about Crossover Dogs

One of my pet peeves is training videos made with dogs who already know the behavior. That is not exactly the case here, but I do realize that Clara has a jump start compared to a lot of dogs, my own crossover dog included.

If your dog is coming from a training background where trying things out has not been encouraged, it will likely take longer for her to catch on to the methods I am describing here, but she probably will eventually. All animals are wired to notice what makes good stuff come their way (as well as what predicts bad stuff). Clara’s readiness to offer behavior is greater than that of dogs who have not been reinforced for that, so naturally things go pretty quickly. I reinforced behaviors I liked from her from the first day she came in my house, without setting up any formal “training.”

But that can also be a good way to start with a crossover dog. You can use some of the dog’s meal calories during the day by tossing him a good treat when he is doing something you like. In her book Plenty in Life is Free, Kathy Sdao describes a protocol she calls SMART x 50, which she suggests as a way to jump start a healthy training relationship with a dog that is not based on a paradigm of making them “work” for everything. When I first got little Zani, who had been cared for but not trained in her previous home, she lit up with excitement when she realized that she could perform behaviors that earned her treats during everyday life.

OK folks, do you capture? Am I the only capturing fiend? I would love to hear how you set about making desired behaviors more likely, when capturing or using any other method.

Related

Eileenanddogs on YouTube

© Eileen Anderson 2015                                                                                                                               eileenanddogs.com

 

Share Button

About eileenanddogs

Passionate amateur dog trainer, writer, and learning theory geek.

Eileen Anderson on Google+

This entry was posted in Clicker, Dog training hints, Operant conditioning and tagged , , , , . Bookmark the permalink.

14 Responses to Dog Training Basics: Getting the Behavior

  1. Natalie says:

    It’s funny. When I first started training, I thought luring was the solution to everything, shaping was kind of cool but usually more effort than it was worth, and capturing was a good thing to know how to do in theory but “really, if you have another option, why would you choose that?” I attached verbal or hand-sign cues to *everything* and gave cues pretty much non-stop to micromanage my dogs, and I called that “well trained.”

    Five years later, I capture probably 75% of the behaviors I teach, and shape virtually all of the rest. I use luring as a temporary crutch to get me to a point where I can use either capturing or shaping. I don’t bother with intentionally adding cues until I actually need them, because so many of the things that I teach have environmental cues and there’s no need to tack a verbal cue on top of a perfectly useful situational cue. I build strong default behaviors, pay attention to environmental factors working for/against me and capture correct responses so that my dog can function on “auto-pilot” a lot of the time and still make correct choices, and now I call that “well trained.”

    Wonder what I’ll be doing in another five years.

    • eileenanddogs says:

      Wow, that’s a great description! That sounds like a recipe for a very happy, smooth life for all concerned. Thanks so much for describing what you do.

  2. Elizabeth says:

    I taught my dog to “bow” using capture, although I didn’t know it was an actual method at the time. I do teach everything with a cue- a gesture I mean. I am a very amateur amateur, so I am still learning all the time. Thanks for sharing your knowledge.

    • eileenanddogs says:

      You are welcome, Elizabeth! A bow is a fun thing to capture. Thanks for sharing about that!

  3. Donna says:

    Thanks for the great videos! I must sheepishly admit I perform much like Natalie describes herself five years ago. But learning all the time and blogs like yours are so helpful! I am relying less and less on the lure and learning more patience to wait and capture. I find your blog so inspiring, thanks for sharing your self critiques, it is very useful for others on the same path.

    • eileenanddogs says:

      You are welcome, Donna! We all have to start somewhere though, don’t we? It is really encouraging to me that people are liking the videos with “warts and all.” I’ll try to do some more.

  4. helenjakeways says:

    Thanks for these videos and post Eileen – the videos particularly are so clear at demonstrating the process of learning and the mistakes are instructive 🙂 Particularly love “completely disappointed puppy” – such a great example of how to inadvertently press the “off” switch!

    • eileenanddogs says:

      Thanks, Helen! That particular session was a hard thing to publish, but I’m very encouraged by the people who have said it helps!

  5. Tania says:

    I LOVE you videos! Thanks for posting them! I also was really inspired about Kathy Sdao’s protocol in her seminar and am trying to use less luring, micromanaging, and am trying to leave the environmental cues speak. It is new to me, but my dog and me are enjoying it so much!

    • eileenanddogs says:

      Thank you! And you are welcome. Sounds like you are making real strides! Environmental cues are cool, aren’t they?

  6. Judy Tricker says:

    Eileen I have just found your videos and they are great. I have clicker trained horses so I have a little idea of how it works and have used the same principle on my American Staffy. The one thing I had to be very careful of was training a behaviour that was a series of his actions when some of them were NOT what I wanted. For example. Every time he came into the kitchen I would tell him to go out and then treat him. This became a bit repetative until I stopped and thought about what I was training and it was this – 1 – Come into the kitchen, 2. Be told to go out of the kitchen 3. Go out as I am suposed to and 3. get my treat. So you can see that coming into the kitchen was part of the thing I was training him into. So then I tried waiting till he was sitting outside the kitchen – before he came in and gave him a treat. Eventually this worked and now he does not get anything if he comes into the kitchen. Only when he is sitting patiently on hhis mat outside the kitchen and not under my feet!

  7. Pingback: Clickers, Capturing Behavior, and Dog Training Tools - The Dog's Domain

  8. Pingback: "Good Sit!" - eileenanddogseileenanddogs

  9. You had briefly alluded to the cue timing in a more recent post, but I came back here to put my comments in context. On capturing, I don’t see it as inherently better or worse that any other approach, but dependent on the dog, person and situation. For any prompted behavior learning by any method, you need cue association and repetition. I feel that if you can accomplish this and have sufficient time available, then capturing is just fine. However, I’ll also agree with Ian Dunbar’s argument that, when time is scarce, lore-and-reward will often accomplish those two things much faster.

    I’ve also seen the argument that since the “captured” behavior is something that the dog himself decided, he’s more motivated to continue it. I feel that’s often nothing more than wishful thinking. An exception which does support that would be shaping from previously habituated behavior.

    As an aside, you referenced Sdao’s “Plenty in Life is Free”. Perhaps an interesting future topic would be directly contrasting that with Dodman’s “Nothing in Life if Free”?

    Now, back to the cue timing during learning. The primary goal there is repetitive association. It might then seem logical that initial association would be strongest if given when the dog is just beginning to perform the behavior. And a subsequent repetition increase would be greatest with the cue as an antecedent prompting the behavior, rather than waiting to capture. And I do feel that is proven in practice, but there’s a bit more to it.

    Some articles on capturing say to keep the cue with the behavior until the behavior is very consistent, but they have offered no basis to support that. We know it will often be quite some time before the association reaches full strength, and it seems to me that, once even a very mild association is formed, from that point onward that repetition may be the more important factor.

    Of course, if you are able to nearly “capture” the behavior 30 times/minute then you have enough repetition that there’s no reason to move the cue earlier until you achieve high reliability. However. if it’s only one repetition every 20 minutes, that argument moves to the other end. Of course, I’m of course using extremes here for the point.

    In practice, after using the cue to mark the behavior a few times, I then try moving it back to an antecedent briefly. If the dog doesn’t respond in a few tries, I’ll start over. But if they do respond (and many do), I’ll shortly have high enough repetition to ensure behavior persistence, and in a far shorter time.

    So, do I use capturing? Sure I do. But I don’t try to use any single method for everything. And many times I’ve heard people say that “this” ALWAYS works so much better than “that”, it’s often because of the particular way they were doing “that” method.

Comments are closed.