Leave it - Better than Before
By Ran Courant-Morgan, M.S., BCBA, CCUI, ADT-IAABC, FPPE
“Leave it” used to be one of my favorite things to teach because it seemed so effective so quickly. I would start by standing in plain view of my dog, who, in most cases, had been eating treats from my hand just moments before. I would then pick up a treat and place it in my palm, closing my fingers around it. And then I would lower my closed fist to the dog’s nose.
Of course, this sweet animal – who, again, had been eating this exact treat from this exact hand just a few minutes ago – would try to eat the treat. They would lean in towards my hand and when my fingers remained wrapped around the treat, the dog might lick at my fingers. Perhaps the dog would lift a paw and smack at my hand, maybe even curling the claws and digging a bit. Sometimes the dog would nibble a little, turning his head to get the best angle, squeezing the teeth in, trying to get the perfect angle to get the treat that I seemed to be offering. After some time had passed with these varied attempts – sometimes mere seconds, sometimes a whole minute – the dog, flummoxed, might stare up at me, or back away – and then I would click and give the dog a treat. “Yes!” I thought I was saying, “Move away, or back up, or do anything except try to get this treat!”
“It worked fast and it was very reinforcing … However, the more dogs I have trained and the more I have learned about behavioral science, the more strongly I feel about avoiding this approach to teaching “leave it” entirely.”
Then we would do it again. On the next repetition, the dog would give up much faster; after just two or three more times, I would hold out my head and the dog would immediately back up or look at me or repeat whatever behavior I had clicked for last time. Then I would gradually start presenting my hand with my fingers open a little bit, until I could hold out a treat in a flat palm and the dog would not move towards it. From there, I could put the treat on the ground, and I could add a cue. “Leave it!” I would chirp, and the dog would back up or look away from the treat.
It worked fast and it was very reinforcing for me. It was also reinforcing for my students, who quickly found success and got to practice their clicker skills. However, the more dogs I have trained and the more I have learned about behavioral science, the more strongly I feel about avoiding this approach to teaching “leave it” entirely. Here are five reasons I have changed how I teach this skill:
First, this method physically hurts. Have you ever had a puppy carefully and systematically try to remove your cuticles because they are in the way of a treat you are holding in your hand? It may be tolerable for some people, but for others it is not, and either way – it is painful. Some dogs just lick a little bit, perhaps nibble once or twice, and then sit back – but others try, and they try long and they try hard. They bite down. They try to pry your hand open one finger at a time. And if you’re doing this exercise “right,” then you need to just sit and wait until they give up, regardless of the bruises and scrapes you may endure.
Second, this method relies first on extinction, and then on reinforcement. The definition of extinction is “withholding reinforcement from a previously reinforced behavior.” In this case, most pet dogs have a history of eating food out of a person’s hand. But in the case of teaching “leave it” this way, we present food in our hand and then, when the dog attempts to eat the food, we withhold that reinforcement. Extinction procedures typically result in what is known as an “extinction burst.” An extinction burst consists of an increase in the magnitude of the behavior (this might look like a dog who was licking and is now nibbling), an increase in variability of the behavior (the dog who was using her mouth is now pawing at the closed fist), and may include an increase in aggression (we might see the dog biting harder or barking). Many dog trainers talk about this in terms of frustration as a “function” of extinction. Regardless, in these moments, the dog is likely not having fun – and neither are we (see previous point on pain).
“Many dog trainers talk about this in terms of frustration as a “function” of extinction. Regardless, in these moments, the dog is likely not having fun – and neither are we. ”
Third, as the procedure progresses to an open hand, it incorporates punishment. Punishment is defined as any intervention that immediately follows a behavior, resulting in a decrease in the future frequency of that behavior. As the dog gets better at leaving food in a closed hand, the person starts to gradually open their palm. Many dogs will continue to turn away or back away from the food in the open hand, but many other dogs will see the opening palm and move towards the food (again, because this is something that has been heavily reinforced in the past for most pet dogs). The person is instructed that, if their dog moves towards the open palm, then the person should immediately close their hand. They are, therefore, removing access to the food contingent on the dog approaching the food – a clear negative punishment procedure. Admittedly, this may be a relatively innocuous punishment procedure compared to something like a reprimand or a physical “correction”, but it is still punishment, which has additional side effects and risks.
Fourth, generalizing this skill can be tricky! Using this procedure, we are teaching a dog that, when they see a treat in a hand or on the ground and we say “leave it!”, the dog should back away or move away from that thing. The humans are very quick to generalize this, asking dogs to “leave it” when dogs show interest in chicken bones on the sidewalk, dogs passing by, or even people. The response we have taught the dog following the cue “leave it” is turning away or stepping away from the thing – and that’s just not an easy thing to do when our dog is looking at a person who is approaching, discarded food or wrappers on the sidewalk just ahead, or a dog across the street. Additionally, having taught “leave it” using extinction and punishment procedures means that the cue has likely been paired with somewhat aversive contingencies along with not being able to access something of interest - even if we’re providing edible reinforcement after the fact. Therefore, using this cue in real life situations could elicit additional stress responses on top of any frustration from a dog who is attempting to approach something that their person doesn’t want them to approach. In brief, it may just not work and, even if it does, it probably doesn’t feel great.
And fifth: on top of all of that, to put it simply, we can do better! What if we think of “leave it” instead as a really solid orientation cue? This is something that we can teach with positive reinforcement alone, that can be fun and positive, and can still be applied to the same situations where someone would use “leave it” as described above! With this method, “leave it” might mean “orient to me.” This is a quick and easy behavior that a dog can do from a number of positions, and for which we can build a strong reinforcement history. Once we have a snappy orienting behavior on cue, then we can start to fade in our distractions. This might first be a food in a closed hand – but out of reach of the dog so that the dog can never get the answer wrong. There are a number of ways to do this; one way that I really like is described below.
Another way:
I like to start with the Up Down game from Control Unleashed (McDevitt, 2019). This pattern game involves standing and facing the dog, putting a treat down on the ground between my feet and the dog, and then standing up again. When the dog finishes eating, she typically looks up towards me, and I click and put another treat on the ground. We repeat this until we have a nice pattern of the dog eating the treat and then looking right up at me.
Once that pattern is established, I can add the cue. Once the dog has finished eating and before she looks up, I say “leave it!” That way, the cue “leave it” immediately precedes the behavior (orienting towards me). After just a couple of repetitions of that, we can add in the item that the dog should be “leaving.” I hold my (empty) fist out to the side at shoulder height. The dog might glance at my hand, and this is when I say “leave it!” She will typically continue to engage in the pattern and orient toward me. Now she is looking at me in the presence of the hand held out to the side.
From there, I will add food to the hand; then hold the palm open; then gradually lower the palm closer and closer to the ground. This will typically not happen over one session, because we want to keep training sessions short (1-3 minutes). Over a few sessions, I am able to place something on the floor, and when I say “leave it!” the dog orients to me. Notably, she orients quickly and joyfully. With this method, she is quickly getting the answer right, and she is rarely if ever encountering any sort of punishment procedure.
“Over a few sessions, I am able to place something on the floor, and when I say “leave it!” the dog orients to me. Notably, she orients quickly and joyfully. With this method, she is quickly getting the answer right, and she is rarely if ever encountering any sort of punishment procedure.”
If she does lunge forward to the open hand, we can take a metaphorical step back in our training. This might involve taking a break, and then in our next session, practicing at the last successful level. Depending on when the “error” occurred, we might also be able to slice the behavior into smaller approximations.
As with any skill we teach, we do need to be mindful about how we generalize, and we need to practice it with a variety of items and contexts – but that is a post for another time (or, better yet, something to work on with an individual trainer!).
I love this method because we get quick results, it progresses quickly, and no one gets hurt! There is limited frustration. We can see success before we ever add food to the picture – and this also means we can wait to add food until we are seeing the responses we want. We entirely avoid extinction and punishment, and we are teaching a specific behavior that the dog can do in a variety of settings.
There are lots of ways that we can teach lots of behaviors. If we look just at what we want the final outcome to be, then either of these methods could be successful - they can both result in the dog turning away from something on cue. However, the way we get there does matter, both for the final behavior as well as for our dog’s experience in training and in moving through the world with us. Using positive reinforcement and avoiding extinction and punishment can help us to build trust (predictable contingencies) and strengthen our relationship with our dogs.
Check out this video of one of our clients, Frog, learning Leave It! This video is comprised of Zoom recordings from two live coaching sessions. Thanks for letting us share this, Logan and Frog!
Thank you to Andrea Viveiros of Connect the Dogs for all of the conversation around the nerdy details of Leave It!