Austen posted something on Insta several months ago about the futility of distracting big dogs in city parks when they get hyperaroused. It made me think pretty critically about my own training paradigm. I didn’t totally agree with her premise because I’ve known a lot of really quality dog trainers to successfully train dogs to walk quietly and calmly on leash around things that used to simulate the CRAP out of them. But on the other hand I kinda really agreed* because I’ve seen a lot of people (and been the person) frantically trying to stuff my their animal’s mouth with treats in a desperate attempt to “distract” them.
(*edited for clarity after I posted this)
if at first you don’t succeed in training your dog to walk on a loose leash
give up and just take her places she can be off-leash
Both counter conditioning and distraction have been major parts of my clicker training program with the Murr-man this last year. And they will definitely both have their place in my training program. But the two aren’t exactly mutually exclusive, and I think that a lot of people accidentally blur the lines between them. This leads to poorly trained animals (and people), and at least one of the misconceptions around clicker training with horses.
Distraction is directing an individual’s attention away from something else (loosely adapted from here — not actually a training term, I don’t think).
Counter-conditioning is training based in classical conditioning that attempts to replace a bad or unpleasant behavioral response with more adaptive or desired behavioral responses (adapted from this definition).
This Facebook post from my friend, a trainer whom I respect deeply for her work with many species and some seriously challenging training tasks (train a giraffe to hold her foot on a radiograph place for ringbone rads, anyone?), added to this thought process and really clarified things for me.
I was like Oh. Shit. The piece I’ve been missing this whole time is high reinforcement for incompatible activities.
To run with the dog example a little more: when I first started R+ training with animals, a friend of mine had a dachshund who barked at nothing, for no reason. Pretty typical dachshund behavior, but this particular dog’s barking was a little more off-the-charts than the other dogs. And I think it was shrill and irritating.
My training approach to this particular behavior was to interrupt the behavior (yelling, stomping, picking the dog up and moving her to a different place) and then start a training session. Which didn’t work. And because I never trained a “quiet” behavior or specifically focused on incompatible behaviors, potentially even encouraged her to bark more. When she barked, I grabbed her and stuffed treats in her face. What’s not to like? In contrast, my trainer friend allows her dogs to give off a couple of alarms and then rewards them for longer bursts of incompatible behavior, like settling on their bed or laying on the floor with her. The dog I mentioned above would never have been able to voluntarily do either of those things, because she was too over-stimulated to think about laying down, unless I physically picked her up and moved her.
i give him a 10/10 for settling and laying down on the — oh crap I don’t want him laying down out here
There’s a superb video of the late Dr. Sophia Yin training a dog to change his association with a stimulus he really hated. This also informed the way I thought about counter-conditioning. Your horse doesn’t have to go from 100 to 0 in order to earn a reward. Any movement of the needle over toward the behavior you want is good. And sometimes — as you see Dr. Yin doing in the video — you do just “distract” the training subject. Dr. Yin also states one concern people have about treating during aggressive events, and a major reason that many people don’t clicker train their horses: they are worried that they will be reinforcing the aggression (or in the case of horses, rude behavior).
(I just watched the video through again and have to profess my love of Dr. Yin. The world lost a bright light when she died.)
scrolling through old pictures I found this one of Murray treating himself to his neighbor’s personal space. because why not?
To bring this back around to Murray, understanding the incompatible behavior part of counter conditioning made a huge difference in our tacking-up training. For Murray, tacking up went like this.
- dislike the girth, get roller-skate-y and fling head in air
- skitter/run away
- girth goes away, yay! instant reward
- if not 3, then do more of 2 and/or other things until 3 is achieved
To change his association with tacking up, I tackled the problem from both ends. I rewarded him heavily for a behavior incompatible with running away that is also a common physiological marker of relaxation: putting his head down. (I learned this trick from my dressage trainer a long time ago — to kindof fake it til we make it with the head-down thing. It will probably take more to explain than I want to in this post, so perhaps I’ll get back to it some day.) Murray couldn’t relax enough to get his head below the point of his shoulder at first, but I just kept at it day after day. I kept the girth stimulus mostly below threshold, and then rewarded him heavily for putting his head down as low as he could get it. Now, if I mess up and Murray gets a little anxious while we’re tacking up I keep doing whatever it is I’m doing (holding the girth on his belly or doing up one buckle, for example) and then calmly wait until he settles and puts his head down low. Then I click and reward him.
The same thing goes for giraffing his way around new stuff. I know he’s scared of this stuff. A carefully misplaced leaf scares him. And that’s okay. But instead of patting and soothing and treating Murray for looking at that scary stuff (potentially valid options if he were a different creature), I ask him to perform an incompatible behavior and reward him for doing that. Usually it’s a low-placed hand target, but sometimes I’ll ask him to take a step or two forward with me toward the stuff and reward him a lot for that.
On the other hand, if I were just distracting Murray I’d be letting him do his own thing (more realistically tugging on the reins and trying to get him to walk with me) and simultaneously stuffing carrots in his face. Which would be pointedly not helping the situation, since it is rewarding behaviors I don’t want (ignoring me, staring at scary stuff without moving, being alert) without asking him to give me any of those behaviors I do want (listening to me, putting his head down, walking past scary stuff).
excuse me you want me to do what?!
It’s a fine line, and counter conditioning and distraction aren’t mutually exclusive. And I’ll absolutely use distraction to break through the haze of NOPE that Murray sometimes gets when we’ve grossly surpassed his stimulus threshold. But I do think it’s an important distinction for us to understand, since as riders we are effectively training an animal (or animals) every day.
Distracting a horse without a plan or any consideration of the desired outcome of the behavior may very well lead to worse behaviors. When I was standing around shoving a steady stream of carrots into my horse’s mouth for just tolerating the farrier, I wasn’t paying attention to Murray’s body language or what he was doing to the farrier. I could easily have been rewarding him for subtly kneeing her in the guts.
On the other hand, a carefully thought out counter-conditioning strategy has worked very well for us in a number of areas.
Any thoughts or additions to this? Was this something I was really late to realizing? My understanding of training is constantly evolving, so I’m sure there is a lot that I’m missing here, and a lot for me to learn still!
Zen and the Art of Baby Horse Management 
as far as i can tell, animals don’t understand what “good” is unless we tell them. if we only ever tell them when they’re bad, how are they supposed to know what to change? like, if my horse is leading quietly in or out of the field? i tell him he’s a good boy, and he’s like, “well how about that, i guess i AM a good boy.” then when he’s being spooky or distracted or pushy or whatever and gets in trouble, he has a baseline of how to be good again.
i think we can definitely over complicate a lot of this stuff, bc it all basically boils down to consistency and clarity in communication. if one behavior elicits a positive reward every single time, no exceptions, and another behavior receives a correction every single time, no exceptions, an animal will figure it out.
LikeLike
I’m fascinated by the barking dog example because I have a Jack Russell. AKA we bark. at. every. thing. And when we get in that headspace, I cease to exist in the world. I don’t think I totally understand this, but I’ll have to go back and re-read it when I’m not at work and see if things clear up.
LikeLike
Interesting. In working with children we always talk about how if you want a behaviour to go away you have to have a ‘replacement’ behaviour. Very similar concepts.
LikeLike
This was a really useful thing to read through, thank you. I don’t have any definitive thoughts but one thing I’ve struggled with on the counter-conditioning front is that rewarding for the good behavior amidst the bad behavior, with smart and fast animals, sometimes just amps up the bad behavior-good behavior cycle. An example: Tristan paws, so I worked for a time on marking the moment when his foot was resting on the ground and then rewarding it. Once he got the hang of that he would paw wildly three or four times, plant his foot firmly, and look at me expectantly. I’m not sure if my timing was screwy or if I wasn’t “building” the good behavior enough by pushing through to the other side. My dog does something similar – loses her tiny mind in flailing at something new, recalls to me and waits for a treat, and then instantly goes back to losing her mind. Lather, rinse, repeat. I’ve thought a lot about whether it’s a question of my timing, my not recognizing and building on the positive moments, or whether I’m not entirely grasping the totality of the exercise.
LikeLiked by 1 person
This totally happens to me too.
Part of what I think you’re experiencing is the extinction burst. This is a well known phenomenon when training the cessation of a behavior. Because the behavior used to be linked with reward (even if it’s just an intrinsic reward, like “yay look I’m pawing I’m so clever”) animals tend to try it EVEN HARDER for a time to see if they can still get the reward. And then maybe the “stop” reward as well. I’m still working try Murray’s extinction burst of pawing. (He does it when he anticipates treats and it’s very self reinforcing because he does it around feeding time and then magically food appears – thanks to his leg magic, I guess.)
The other thing that has helped has been switching to a lower value treat. For new behaviors I use very high value treats to get a lot of offered behaviors that I can then shape. But when I want Murray to stand quietly, I use low value treats so he doesn’t get too excited.
Just some thoughts – I’m by no means a great trainer!
LikeLike
It’s really hard to go in depth in training theory in an insta post, but I tried to hit on the most annoying things. I use distraction all the time in training. My dogs are huskies, prone to hyper focus. I break their concentration constantly. The difference is in redirection and the ability to have other commands to rely on to bring relaxation and a quiet listening mind back to the forefront. For example, I distract my dog and run through a series of sit, down, roll over commands which get her using her brain and listening. She stops acting on instinct and starts thinking about the situation. I do not distract dogs with food, in my experience this only serves to amp them up further. I am also absurdly clear with good/bad behavior. My dogs know when they’ve done right, and when they’ve done wrong. They’re smart enough to try the list of appropriate behavior when told they are in the wrong. But they must know clearly what is appropriate to do so.
With horses things are slightly different. I will distract with food with horses, for example. I think food helps a horse relax, unlike a dog. But, the premise is the same. I think we’ve all seen a fractious and worried young horse settle when put to work in a way he’s familiar and comfortable with (like how most horses settle in a spooky environment midway through their dressage test). That’s what I strive to achieve. It’s a multi step process as you must first develop the historical working relationship you’ll need to rely on in the stressful trying situation.
LikeLiked by 1 person
I did a terrible job of saying this, but I actually REALLY agreed with you. By the time your dog (or horse) is dragging you across the road at something, the time to distract them had come and gone. Your post really made me think more carefully about how I use both in my own training.
So much is lost even in a longer (poorly written, late-night) post. Your comment pointed out another really, really important factor: reinforcement history. Murray now knows that there are set parameters of behavior that are rewarded, as opposed to before clicker training when carrot dispensing was essentially random. That consistency has done tons of the work for us. And it means that my training efforts everywhere are more effective than they were before, because he knows what the “game”/ expectations are.
LikeLike
This is absolutely fascinating, at least partly because I’m pretty clueless when it comes to this stuff. I spend 99% of my time with Frankie scratching him and petting him, and when he does something bratty I smack/growl at him. And then we go back to scratches and pats when he stops. That being said- he came to me with ground_manners.exe already installed and updated, and I haven’t had to do anything beyond simple reinforcement. Definitely enjoying learning more about how those behaviors are learned and tested!
LikeLike
Interesting!!
It’s kind of neat how we as people, end up conditioning our dogs to our responses (for both good and bad), sometimes without even realizing it. What I’ve learned about dog (and horse) training is that your timing has to be very accurate and without that, you end up with an animal that escalates into the black hole of no return and it takes quite a bit to fish them back out again.
LikeLike