Hi, is it bad to do a negative reinforcement when the animal doesn't do the desired behavior, if it is verbal? For example, is a strong verbal "no" harmful to the animal or the teaching?
Thank you for the ask!
First of all, that’s not an example of negative reinforcement, so I’ll do a quick breakdown of the types of operant conditioning :)
Positive  reinforcement is a consequence that occurs during or immediately after a behaviour that increases the chances of the animal performing that behaviour again. This is ethically the best type of conditioning that should be used during training, because by adding something the animal likes and will work for, you give it control through choice, because it does not have to participate if it doesn’t want to. Example: giving treats, attention, affection, to a dog for sitting; giving toys, stickers, or praise to a child for doing their chores; saying thank you to someone for doing any nice thing (this is, coincidentally, why I say thank you in response to every question I get here!)
Negative  reinforcement is also a consequence that occurs during or immediately after a behaviour that increases the chances of the animal performing that behaviour again. (Basically, any type of reinforcement aims to increase the likelihood of a certain behaviour occurring again). The difference is that, instead of giving or adding something the animal will work for, you are removing something it will work to avoid. This still gives the animal some choice, but animals trained with negative reinforcement will usually only operate at the level necessary to avoid the negative reinforcement stimulus. Example: the annoying noise your car makes until you put your seatbelt on (putting it on is reinforced by taking the annoying noise away); yelling at an animal to go to a certain spot.
Positive punishment is a consequence that occurs after a behaviour, and is intended to decrease the frequency or intensity of that behaviour by adding/giving something undesirable or perceived as harmful. Classic example is hitting an animal because you want it to stop barking, biting, etc.
Negative punishment also has the purpose of decreasing the chance of a certain behaviour being performed, but by taking something away from the animal that it enjoys/needs. This includes things like starving an animal for doing something “bad”, or taking away video game privileges from a child for being rude, etc.
Your example (a strong “NO”) sort of blurs the lines between negative reinforcement, if you think it’s going to make the animal more likely to do the behaviour you want (which it absolutely will not), and positive punishment, if you’re trying to eliminate other behaviours that you don’t want. No professional/contemporary/ethical animal training company/facility uses punishment for the following reasons:
The animal isn’t given any control.Â
This method has  no  predictable  outcome  because  it  does  not  teach  the  animal  what  it  should  do, which is also just inefficient.
The  act  of  punishing  is  often  reinforcing  to  the  punisher. Those who rely on nothing but punishment, whether it’s with human or animal relationships, will only continue to do so because they think the punishment is working, and because it gives them a sense of control, power, etc.
Many animals do  not even understand  punishment. Especially if the punishment is horribly timed - for example, yelling at your dog when you come home and discover hours later that it pooped on the carpet. The dog is not going to make the connection between what it did hours ago, and you yelling at it now.
It  can  produce  serious  side  effects, like depressive, anxious, avoidant, and aggressive behaviours.
In short, unless the animal you’re training happens to be another human, non-human primate, parrot, or any species that explicitly understand human language, a strong verbal “NO” is essentially useless, because it will simply not understand you. Aside from that, there might be no harm to the animal as long as you’re not yelling at it and causing fear and stress, but there is harm to the training process by slowing it down and not making any progress at all. Punishment and negative reinforcement may seem like the easiest way to get what you want from an animal, and it is what humans usually turn to first, but positive reinforcement is the most beneficial to both the animal and trainer. It is also the more efficient way, if you’re willing to put in the small amount of time and effort at first to learn to be empathetic and try to imagine how the animal perceives you and what you’re asking from it.
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!
