This year’s Art and Science of Animal Training Conference seemed to focus on a lot of the issues which are a little controversial in the positive reinforcement community right now, such as…
- CAT-H (Operant Counterconditioning)
- The potential for coercion in Positive Reinforcement
- That Negative Reinforcement cues don’t necessarily function as conditioned aversives
- The idea of clicker compatible Negative Reinforcement
Katie Bartlett attended the conference and has done several write-ups on the talks, which I highly recommend, and I’m going to quote her here…
“Negative reinforcement is the elephant in the room. Everyone is trying to ignore it, but that’s not really a solution. Perhaps it would be better to learn to recognize it and become more educated about it. Then we can decide if we do need to throw it out or if we can modify it so that it is compatible with clicker training.
What does this mean? First, it means that we aren’t abandoning the use of lead ropes and reins. They are useful tools and necessary when we take our animals out and about. Think about all those dogs on leashes at ClickerExpo. That is accepted as a safe and normal way to take dogs out in public. So it’s not that leashes are inherently bad. Instead, the critical question is:
Does a lead represent cues or commands?
To answer this we have to look at the teaching process. That is what matters. A good example is the technique (Alexandra Kurland) calls ‘Shaping on a point of contact.’”
– Katie Bartlett
As a trainer with traditional clients, even though I love using positive reinforcement, I obviously mix with traditional pressure and release/negative reinforcement as well.
It’s important to understand that I use negative reinforcement not because I don’t know how not to. I actually spent several years training with pure positive reinforcement, so I know how to build behaviors without negative reinforcement, but I CHOOSE to use it because I feel like it has a place with the animals who I work with.
Generally, I like to teach with positive reinforcement, then transfer everything over to more traditional cues. I feel like this allows me to fast track a horse quite quickly so I can provide them as broad a foundation as I can, while prioritizing calmness and confidence.
Everyone’s implementation is going to look a little different, and there may be trainers who choose not to use negative reinforcement all, or who choose not to use positive reinforcement at all.
I’m not here to argue for one way or another.
I’m here to provide more opportunities for discussion, because we need to be able to talk about these things, because we’re all in it for the horse, no matter how we choose to serve them.
Whatever our niche in the industry, we should all be prioritizing setting the horse up for success in each individual situation.
The important thing is, no matter how we choose to train, we are honest with ourselves about what is really happening, that we don’t try to sugarcoat how the animal is actually experiencing our technique.
This also requires that we are objective in our observations about how the animal feels about something which we might’ve been taught is ’wrong,’ like mixing.
It should become clear that the communication and connection that we share with horses through physical touch is nuanced, and cannot really be compared to shocking rats in a box.
I see a lot of trainers getting caught in ‘paralysis of analysis,’ and having some pretty heated discussions about learning theory.
But the problem is, it’s just that, learning theory. When we go from theory to application, no matter how black-and-white things seem to look on paper, once we start applying something to a real animal, we realize that things are a lot more complicated and nuanced.
This complexity can be really disconcerting to those who feel more comfortable with the idea that there is a simple black and white cookie cutter approach that can be applied indiscriminately. Encapsulation within our niche and the resulting myopia is an issue on both ends of the spectrum, from positive reinforcement, to traditional.
As of right now, I have several different ways I think about negative reinforcement in my training.
This understanding and application may change over time.
CLICKER COMPATIBLE TACTILE CUES
We might choose to teach something with positive reinforcement, and maintain that behavior with positive reinforcement, but add a tactile cue that ‘looks’ traditional. Technically, this is NOT negative reinforcement.
Real-life example: when I’m ‘halter-breaking,’ I’ll start out with a hand target, and then that hand on the lead transfers the tactile cue.
CLICKER COMPATIBLE NON-ESCALATING PRESSURE
Alexandra Kurland calls this ‘Shaping on a Point of Contact,’ and it’s where we apply contact, and then wait and shape with positive reinforcement.
Here, contact functions as a cue to earn a reinforcer. To quote Bartlett: “The lead is a source of clues not threats,” and the horses “learn to recognize contact as useful information that leads to reinforcement.”
Real-life example: when I’m ‘halter-breaking,’ I will put a contact on the lead, and wait, and then mark and reinforce when I see something I can build on, like the nose tipping, or a weight shift, etc.
I’m not convinced this is technically negative reinforcement either. The behavior is not maintained by the contact, but by the positive reinforcement. I think of this as using contact as a clue.
In dog training, this is referred to as ‘molding,’ but for some reason, equine trainers haven’t seemed to adopt it, and keep referring to it as an aversive.
Some people derogatorily refer to this as ’negative reinforcement with a cherry on top,’ I’d rather prefer the term. I like cherries.
NON-ESCALATING PRESSURE AND RELEASE (-R)
We can also forgo the positive reinforcement here, and just use this ‘contact and wait’ in our pressure and release/negative reinforcement applications.
Good horseman have been doing this a long time, including the Dorrance brothers and Ray Hunt.
It was all about letting the animal work his way through the puzzle, instead of us escalating the situation and demanding a response through some misguided application of ‘dominance’ or ‘respect.’
Real-life example: when I’m ‘halter-breaking,’ I will put a contact on the lead, and wait, and then release when I see something I can build on, like the nose tipping, or a weight shift, etc.
Real-life example: when I’m ‘halter-breaking,’ I will put a contact on the lead, and wait; but if the animal isn’t ’getting it,’ and is confused, I might choose to reduce mental stress by adding a little more feel to clarify my intention, maybe off to the side so the weight starts to shift and become unbalanced, then I can mark and reinforce or release.
Here’s where we get into tricky territory, because once we start escalating, we might think it’s easier to just bull our way through a problem, instead of waiting and letting the animal think their way through the puzzle.
The fallout of continually escalating can be an issue, as well. When we are filling the bucket of pressure and aversives more than we are filling the buckets of release or positives, the imbalance can wreck havoc on our relationship.
The question we have to ask ourselves is, are we escalating in a way that brings clarity and reduces mental stress, and which doesn’t cause harm, or or are we just escalating to bull our way through, with no regard to how we are affecting the animal emotionally or physically?
Good training requires discernment.
These are difficult questions to ask ourselves and discuss amongst ourselves, in both the positive reinforcement community as well as in the traditional horsemanship community, but it’s important we all feel like we can talk about them, so that we can find answers that best serve the horse.