This post is outdated (I can’t believe I wrote it four years ago now) and I’ll be the first to admit that some of the things I said are wrong and that I’ve learned a lot since then and changed the way I think about and use the different types of operant conditioning. The one I really want to talk about today is what’s called the “constructional approach” to animal training/behaviour modification, which I came across while renewing my Avian Fear Free Veterinary Certification (course created by world-renowned animal trainer Barbara Heidenreich)!
Yes, positive reinforcement is a great tool and what we’d ideally like to use in most training scenarios. However, it’s contingent on the animal’s willingness to accept reinforcers (treats, toys, attention/affection, etc) from their trainer, and the reality is that this just isn’t always possible. This is usually the case for animals who are “over threshold” (i.e. experiencing very high levels or fear, anxiety, and/or stress). For example, a parrot in a new, stressful environment such as a vet clinic may be too stressed to accept treats or scritches even from its owner, and similarly, a fearful-aggressive dog with a history of being abused probably won’t be keen on being approached with snacks and pats right off the bat. This is where careful application of negative reinforcement can actually be used quite effectively. As a refresher, negative reinforcement involves the removal of an aversive or undesired stimulus in order to increase the likelihood of a certain behaviour being performed. Let’s break down the components of negative reinforcement in the case of an animal who’s over threshold:
The behaviour that we’d like to see more of/that’s being reinforced is calm, relaxed behaviour.
The aversive stimulus is the presence of the trainer/whoever wants to get close to the animal – this is what’s causing fear/anxiety/stress.
So, we can use distance from the trainer as the negative reinforcer for calm behaviour!
Example that I’ve done in real life: There was a hawk at the raptor center I used to work at who was a new arrival and showed fearful-aggressive behaviour in response to being approached by people he wasn’t familiar with. He would get into a defensive position and crouch, holding his wings out and puffing up his feathers to make himself look bigger and poised to attack. What we wanted to see from him was calm behaviour – feathers in a relaxed position while standing normally or just doing his own thing. Rather than invading his personal space while he was already being defensive to toss food and treats his way in the hopes that he would eventually form a positive association with us and relax, we made the process a lot faster and less stressful by using distance from us as a negative reinforcer. We would observe him to make sure that each session started off with us being far enough away for him to remain calm and relaxed without interrupting whatever he was doing. Then, step by step, we would slowly approach him – but immediately stop as soon as he changed his behaviour and showed any sign of fear or anxiety toward us. Next, we would stay completely still and quiet while waiting for him to shift his attention away from us and perform any other behaviour, even if it was something as small as just looking in a different direction, shifting his weight from one foot onto another, etc. As soon as he did, we would calmly walk away until he couldn’t see us. And we repeated this process over and over again until we were able to stand right in front of him without any signs of his previous aggressive behaviour, at which point we switched over to using positive reinforcement with food rewards to get him to step up, and so on.
The point is that in the beginning, he was able to understand that in order to get what he wanted, which was distance away from the scary thing, all he had to do was remain calm, and we would respect that and leave. When negative reinforcement involves the use of mild aversive stimuli and excellent attention to body language, it can be an effective and ethical method of training an animal and gaining their trust. Previously, animal trainers/behaviour professionals (myself included), avoided using negative reinforcement whenever possible because of its tendency to cause animals to be less enthusiastic to engage in training and even worsen aggression caused by fear/anxiety/stress. And this is definitely something that’s still valid – negative reinforcement is not ok or ethical when the aversive stimulus is significant and not something that already exists in an animal’s environment. For example, tightly squeezing/applying excessive pressure to a towel wrapped around an animal while attempting to restrain them for an exam/other procedure and decreasing pressure only when they stop struggling or “calm down” (this can actually cause chickens, parrots, and other small prey animals to succumb to heart attacks if they are not used to being handled).
I also wrote about my utter disapproval of punishment before, but negative punishment (taking away something that an animal enjoys/desires with the intent of discouraging a certain behaviour) can also be used ethically in some situations, as long as it’s also done very carefully and followed up almost immediately with positive reinforcement. It’s particularly useful for animals exhibiting attention-seeking or playful-aggressive behaviour – parrots, kittens, puppies, etc. who suddenly bite your finger too hard after playfully mouthing at your hand. When this happens, you can punish the biting behaviour by taking away what the animal wants in that moment - playing with you/your hand. Immediately stop playtime and take your hand away (this is when a gentle verbal cue like “no” can be used as a marker, as long as it’s not too loud/aggressive so as to startle and frighten the animal), keeping in mind that this “time-out” period should only last a few seconds. Then, quickly switch to positive reinforcement, either by asking the animal to perform a behaviour they already know (e.g. step up) and offering a reward, or by waiting for them to demonstrate calmer behaviour (nudging/licking your hand) and reinforcing that by resuming play, giving attention/treats, etc. This provides clear communication to the animal about what behaviour is and isn’t acceptable in order for them to get what they want.
I could go on and on about lots of other examples and more (like how even positive reinforcement can have a “dark side” when used in a coercive manner…but I think this is a good enough update for now!