07/12/2023
THE DARK SIDE OF POSITIVE REINFORCEMENT…
‘FAUX COUNTERCONDITIONING’
Over the years, I’ve had several different horses come in for training, who, in spite of extensive positive reinforcement training, were still struggling with a lot of anxiety in their day-to-day lives.
By no means was this due to a lack of quality training.
These horses were ‘performing well’ within the context of a formal session, some with very experienced and educated trainers.
But outside of that context, they were reverting to almost feral behavior, and highly reactive, even with the same stimuli.
After years of training, they were plateaued in their ability to navigate the domestic world safely and be handled safely.
The assumption would be that their reactivity was a response to being ‘over threshold,’ or not generalizing, but in my search for answers for these horses, and in visiting with other trainers, including the dog training community, I became familiar with the idea of ‘overshadowing.’
The animal isn’t merely over threshold, or not generalizing, they were never OK with things in the first place.
Zoos are often used as the poster child for the implementation of successful positive reinforcement, but it’s Barbara Heidenreich’s work in particular, with ‘positive reinforcement gone wrong’ in zoological settings, that really opened my eyes.
Sometimes what we think is counterconditioning, is actually overshadowing, meaning the food reinforcer and associated behavior are masking the animal’s true feelings and behavior around a stimulus.
If the context changes, or the food reinforcer isn’t present, the animal shows us how they really feel about that stimulus.
A good example is counterconditioning haltering.
I use this all the time, and it works great, and I know it’s working if I can halter without the food reinforcer present, because it means the horse actually has a neutral or positive association with the halter.
If the food reinforcer isn’t present, and the horse shows a fear response, I know that I was merely overshadowing/masking their true feelings about the haltering process.
There’s a lot of discussion about learned helplessness in relation to coercive training and pressure and release, but overshadowing is just as prevalent.
Make no mistake, this happens even with top R+ trainers in the industry, but there’s a stigma about discussing it openly, and admitting that positive reinforcement can go wrong.
Positive reinforcement is just as vulnerable to human error as any other form of training.
There is a learning curve, and novices and even accomplished trainers can make mistakes that are dangerous to animal and handler.
What’s frustrating is that the pro trainers and amateurs who’ve reached out to me about the accidents and injuries they’ve had due to overshadowing, they don’t feel safe having a public discussion about.
I think there’s a fear that admitting to these accidents and injuries will somehow hurt the acceptance of positive reinforcement in the mainstream, or that they’ll be attacked by their peers.
This lack of transparency is not just hypocritical, but dangerous.
The truth is, the only way positive reinforcement is going to go mainstream is if we can be transparent about it.
Like any approach, checking in with our animals, and making sure our training is cognitive and emotional, and not just behavioral, is incredibly important.
I think this article does a great job of explaining some of the things myself and others have observed with overshadowing, as well as how to address it…
https://thecognitivecanine.com/when-well-executed-counterconditioning-plans-fail/
“Talk to ten qualified trainers about their experiences using counterconditioning and desensitization (CC/D) in the field and you’ll likely get ten different responses.
But what might surprise you is how many of them will somewhat bashfully admit that these protocols fail just about as often as they succeed–if not more so.
Less-experienced but well-read trainers will scoff and assume the protocol wasn’t carried out well; that pieces were missing, steps were skipped, etc.
Certainly this is the case some of the time, but what about when a well-designed expertly-executed CC/D protocol fails? Why is this happening?
Something *is* flawed. It may not the plan in the traditional sense, or the ex*****on.
It might be that our currently held information about CC/D is what is slightly off, and it might be that traditional CC/D protocols are less-capable than we thought…”
By the way, that’s me in the picture, shaping forward movement off a hand target with a ‘reverse round pen.’
I love training with positive reinforcement!
I’m NOT anti-positive reinforcement, and I spent several years using positive reinforcement exclusively, no pressure and release at all.
Every behavior I train now, I know I can train with positive reinforcement, but I choose to mix to best adapt to each individual and situation.
NOTE: If you plan on a reverse round pen setup like this, make sure you use rails from cone to cone that tip over and NOT rope as it’s a drag hazard.