Recently, I’ve been listening to a new podcast. It’s the Equiosity podcast, produced by Alexandra Kurland and Dominique Day. The conversations that Alexandra and Dominique have on the podcast center around horse training. However, the ideas they discuss are usually much bigger and apply to working with any species of animal. I think you would enjoy the podcast even if you are not a “horse person.”
In several recent episodes, Alexandra and Dominique have discussed poisoned cues. So, I’ve been thinking some about poisoned cues this month. I wanted to share an example in this post of something in my life that I recently realized had become a poisoned cue.
A quick note: I’m not going to go in depth in this post regarding the definition of a poisoned cue. If you are unfamiliar with the term, I recommend that you check out this post on my blog which discusses poisoned cues or this recent Equiosity episode about poisoned cues. Also, I would encourage you to sign up for the webinar that Alexandra and Dominique will be doing at the end of the month with Dr. Jesús Rosales-Ruiz. They will be discussing poisoned cues, loopy training, movement cycles, and much more. It will be a great opportunity if you have questions about poisoned cues.
Spanish apps and poisoned cues
Over the past year or so, I have been learning Spanish. I currently use several different apps to help me practice.
There is one app in particular that I use a lot. It has a handful of different types of questions that alternate randomly. Sometimes you see a Spanish (or English) phase and have to type a translation, sometimes you listen to a phrase and have to type a translation, and sometimes you have to answer multiple choice questions, such as identifying the gender of a noun or selecting the correct conjugation of a verb.
Then, there’s my least favorite type of question. Sometimes, a Spanish phrase appears on the screen and an audio clip plays the phrase. Then, you have to press the microphone button and record what the speaker said. The advantage of this type of question is that it gives you an opportunity to practice speaking phrases and sentences in Spanish.
If I get the question right, the program then gives me a printed message that congratulates me, and I get to move on to the next question. However, sometimes the program instead gives me a pink bar with red text that tells me:
“Hmm…. That doesn’t sound right.
Give it another try.”
So, I speak the phrase into the microphone again. And the program then almost always gives me the following message:
“That still doesn’t sound right.
Give it one more try.”
So, I try again. If my pronunciation is still incorrect, the program tells me in bright red text:
“Let’s move on from this one for now.”
Let me tell you one more thing about this type of question. When the program gives me this type of question, there’s always a little grey button at the bottom that says “I can’t speak now.” I can press this button if I want to skip the question.
This button is there because in some places, I may not want to attempt this type of question. For example, if I am in a really noisy setting the microphone won’t be able to clearly hear what I say.
However, I’ve found myself more and more often hitting the “I can’t speak now” button when I reach this type of question, even when I’m home alone sitting in my living room! What I’ve recently realized is that I am actually avoiding this type of question. This type of question has become a poisoned cue.
Why would I avoid these questions?
Even though I only sometimes get these questions wrong, I find the feedback I receive for the wrong answers incredibly frustrating and aversive. The program never tells me why my pronunciation is wrong, it just gives me a red message and makes me try again. Almost always, I then get the pronunciation wrong the second time and the third time. It’s a relief when I finally get the message telling me I can move on, even though my answer was never right.
I know that it’s good for me to practice my pronunciation. And, it’s only a little bar of red text. How bad could that be? Yet, it’s become incredibly aversive to the point that I would rather just avoid these questions. When I see one of these questions pop up, I don’t know what the outcome will be. I might get the question right. Or, I might have to go through multiple repetitions of trying, being corrected, and trying again, without ever really knowing what I am doing wrong.
In a recent podcast episode, Alexandra mentioned that ambiguity helps create poisoned cues. For example, a response is sometimes followed by positive reinforcement. However, it is sometimes followed by a correction or something else the animal finds unpleasant. And, the animal is not ever completely sure whether it will be one or the other. This ambiguity leads to poisoned cues and can create long-lasting emotional effects.
I think it’s this ambiguity which has poisoned these questions for me. When I see one of these questions, part of me does want to practice my pronunciation. So, I hesitate as I decide whether or not to give the question a try. However, I already feel frustrated before I even try the question because I don’t know if my answer will be followed by a “good job!” or several frustrating corrections.
One issue with this Spanish app is that it is not really teaching me pronunciation. There is no instruction or progression of shaping steps. Instead, the app just gives me entire sentences, asks me to pronounce them correctly, and corrects me if I’m wrong. This is what animal trainers call “lumping.” It is much too big of a step, and it is often a step that I am not ready for. Lumping often creates frustration. It can also lead to poisoned cues if the teacher is correcting unwanted responses.
I imagine that dogs, horses, and other animals must feel similarly to the way I do with these Spanish questions when we uses corrections or aversives during training.
Imagine you are training a dog to walk on a leash. You give the dog a treat every handful of steps for staying in the right position by your side. However, occasionally, the dog gets just a little bit out of position and you give a strong upward tug on the leash.
From the dog’s perspective, he wants the treats. So, he’s trying to stay by your side. And then, all of a sudden, you jerk hard on the leash. The dog is then trying to figure out what he just did so that he can avoid getting leash pops in the future, but he’s probably not quite sure why he got that one in the first place.
Leash walking practice may very likely become poisoned for this dog. This is because the leash has become ambiguous. When the leash is on and he’s walking with his owner, he will now be unsure if after walking the next few steps he will get a treat or a leash correction. This creates uncertainty and is likely going to make leash walking practice a lot less fun for the dog.
Even once the dog is walking nearly perfectly, he may still be unsure about the situation since leash walking has sometimes produced corrections in the past. And, unlike the skip option in my Spanish app, the dog probably won’t have the luxury to tell his owner that he would rather not work on leash walking anymore.
Trainers sometimes insist that they need both rewards and corrections because “the animal needs to know when he is wrong.” The problem with corrections, however, is that they can create poisoned cues. In addition, they only tell the animal that he is wrong. They don’t help the animal learn what you wanted him to be doing instead.
I believe that our animals are intelligent creatures. If they understood exactly what we wanted them to do, they would do it! If the animal makes an “error” or “mistake” it is because I have not been clear enough in communicating what I want and teaching the animal to do it.
My experience training a variety of species has showed me that we don’t need to use corrections to train. Instead, we can use positive reinforcement and shaping to explain to the animal exactly what we do want him to be doing. If the animal is making a lot of mistakes or is engaging in a lot of behavior that I don’t want, it’s time for me to change my OWN behavior, rather than correcting the unwanted behaviors.
If you found this post interesting, I encourage you to sign up for the Equiosity webinar this weekend with Dr. Jesús Rosales-Ruiz, Alexandra Kurland and Dominique Day. I am sure that it is going to be full of interesting conversations related to poisoned cues and other topics. And, even if you’re not available on Saturday, the webinar will be recorded so that you can listen to it afterward.