I’ve been inspired recently to finally write about something that often happens when we first start to train our horses with positive reinforcement.
It happened to me. It will be happening or have happened to you as you switch from aversive training to training your animals with positive reinforcement.
It will be happening to those people who are sometimes using aversives and sometimes using appetitives (food or scratches) as reinforcers with their horses.
It will be happening to those who are using aversives to get behaviour and then clicking and treating (or just treating their horses) for it.
It will also be happening to people who think they have never intentionally made themselves aversive to, or used aversives with their horses, but who have horses who have been owned by other people who have done so.
You will recognise it when you see the difference between how your horse reacts when you are with him, compared to how he reacts when, say, you have a family member or friend who rarely sees your horse come and visit, or a positive reinforcement instructor come and teach you how to train your horse.
What you will notice is that your horse may react differently to someone she rarely sees, or has never met, or who they have only known as someone who comes bearing treats.
It’s the story of how I watched my horse fall in love with a lovely equine behaviour and training specialist Dr Helen Spence – right in front of my eyes – and what it brought home to me.
It’s pretty sad and at the same time enlightening and empowering if you choose to let it be. I hope you do because my intention is not to make you miserable or guilty or regretful but to actually inspire you to make some really significant changes to how you behave around your horses sooner rather than later.
It’s the story of how easily we can become a source of conflict for our horses and how that can affect their emotions and behaviour.
Whenever I go to teach someone I am aware of it, and I am tuned in to the behaviours that horses will show when they are anxious about the person at the same time as wanting to stay around to get the food. And that is because I see it an awful lot.
The reason I mention it is because a friend has recently become more aware of the phenomenon known as the poisoned cue and has been particularly emotionally affected by seeing how an animal “looks” when they are experiencing this kind of training.
The term, coined by Karen Pryor in 2002 refers to a cue for a behaviour that is associated with both negative and positive reinforcement at the same time.
To poison a cue all it would really require is for a person to give a cue, enforce the cue (make the animal perform the cued behaviour) with some kind of aversive stimulus, and then remove the aversive and click and give the animal a treat.
If the animal were to then perform the behaviour on cue, the trainer would click and treat. If when the cue were used the animal did not perform the behaviour, an aversive stimulus would be applied to compel the animal to perform the cued behaviour, and the trainer would then click and treat.
What happens in this situation is that it produces a different kind of emotional response to that seen when either negative reinforcement alone, or positive reinforcement alone are used.
Animals typically show reluctance, and visible signs of anxiety or distress and very often the animal can appear to freeze or show hesitation in responding.
The advice of those in the know about this phenomenon is that when we realise that we have poisoned a cue, the best thing to do is to completely change that cue. Retrain the behaviour from scratch, using only positive reinforcement (and one of a choice of free shaping, luring or target training), and introduce a new cue altogether. If that is an audible cue it needs to sound totally different, and if it is a visual cue it needs to look completely different.
That means that we need to move differently if our movement is any part of the cue, and we need to make a totally different sound if we are using a vocal cue.
And that’s all very well. But there is a limit to how different we can make those cues and even to how practical it is to do so.
If we are training our horse to lead, for instance, we can’t do that without walking ourselves. And for many horses, the movement the human makes walking is a conditioned predictor of the lead rope pulling on the halter on their head, for any horse with a traditional or natural horsemanship training background.
So the cue may be extremely difficult to change – even if we use a different vocal cue (if we ever did use a vocal cue in the first place) because our movement is an unavoidable part of the cue.
But the killer thing to realise – and it kills you inside when you do – is that so is our very presence there, while all this is happening – at our hands – with the animal.
Everything in the environment can become associated with and therefore part of the cue. And in fact our very presence is a cue in and of itself. When we are with our animals, we are the universal stimulus that predicts both positive AND negative reinforcement – appetitives and aversives – if that’s what we are doing.
It’s important to remember that there are only two forms of reinforcement. If we come from a background where we haven’t been using appetitive training (using food or scratches as reinforcement for behaviour that has not been produced using any kind of aversive stimulus) there really is only one other kind of reinforcement we can have been using if we’ve been intentionally trying to train behaviours and get those on cue.
And that’s negative reinforcement. Behaviour that is trained through negative reinforcement is produced by the application of some kind of aversive stimulus. It’s the cessation of or reduction in the strength of the aversive that reinforces the desired behaviour.
The only other alternatives to getting an animal to do (or not do) what we want, are sedation, manipulation and restraint.
In the end, even if we completely alter how we train our horses – overnight, all-in, removing all use of aversives (all and any kind of pressure, coercion, force, threats or corrections) from our training approach, we simply cannot expect to instantly change how we have come to be perceived.
Not overnight, not in a week, maybe not for years, and perhaps in some cases with severely human-traumatised horses, never.
In the same way that if we started to wear completely different clothes around people, I am not at all sure that even if we wear different clothes, a different hat, change the way we walk or try to use a different tone of voice that any horse is going to be fooled by that.
For one thing it would be immensely difficult to sustain and for another there are way too many things about ourselves that we do or that we “are” that differentiate us from other humans or other animals.
The first time we often see this poisoned cue “effect” is when we first start to train our horse with positive reinforcement.
We can see it either when we are training a default behaviour – stand still and face forward (differential reinforcement for behaviour incompatible with foraging on us) or doing some basic target training, or even during some initial desensitisation and counter conditioning to something – including to ourselves as humans or ourselves as an individual.
What we can see is the horse showing sign of anxiety. They come close because that’s where the food is, but then what we see is them fidgeting or face-pulling or ears back or the turning the head away as an appeasement signal / calming signal. This is happening because the horse feels threatened, but they are staying put because they want the food.
It’s a perfect example of a behaviour performed in a conflict situation where the horse wants one thing – the food – but wants to avoid the other – being chased away or in any way treated aversively – as is their reasonable expectation when their whole life experience has been of mostly aversive handling from people.
What they most often show though is a mix of frustration about the food and anxiety about the situation and the person.
So the nipping, the nose pushing, the behaviour of walking across us when we try to walk with them or away from them, the ears back, the tightness around the face, tail swishing, pawing, fidgeting, distractability, lack of focus, geldings dropping their willy, yawning, licking and chewing or even turning to scratch themselves a lot are so often borne out of that toxic mix of anxiety about us as a species or as individuals tangled up with frustration about not being able to get the food out of us, or to even be able to think straight about how to.
Very often this behaviour is attributed – incorrectly – to clicker training. It’s believed that it’s the clicker training with food that causes horses to nip, push, barge, bite, charge, drop their penis or snap at us.
It’s not. It’s the fact that the horse is in a conflicted emotional state. And that conflict is caused not by the food alone, but by the combination of food and the aversive expectations the animal has because of what we – or others – have done before.
What this means is that when we are training we need to be thinking really hard about how to mitigate and reduce that, as well as recognising (and not beating ourselves up about the fact that) it’s inevitable to some extent because we are aversive predictors when we first start out.
It’s why we so often advocate protected contact – training from behind a barrier. It means that as we learn to change our own behaviour in response to undesirable behaviour from our horses, we are less likely to react reflexively and defensively if a horse goes to nip or push us – because to do that would be a perfect way to confirm their pessimistic perception.
The other reason for protected contact is so that we can have some distance between us and the horse. If the horse is anxious with us close but stays close to us for the food, then it’s very difficult to even begin to help them be calmer. But if we stand a little way away from them – something that the barrier would allow us to do without them following – that makes it impossible for them to perform some of the behaviours we would not want to see, and would find hard to ignore, and it can help them to feel less anxious about what we might do.
The final thing I could add on this is that it’s the big reason why we advocate short sessions of training early on.
These two things – training in protected contact and short sessions – are things that it seems so hard to convince people to do. It’s as if people feel that they aren’t good trainers if they have to be behind a barrier or if they can only train for 1 minute. They aren’t. The best training I’ve seen has been behind a barrier and for 20 seconds at a time.
And the reason for this is that, like any other aversive to which we are trying to desensitise and counter condition the horse, we want to make those exposures really short, and sweet and over fast.
Low strength, short duration exposures to mildly feared stimuli would be the recipe for desensitisation and counter conditioning.
If we are trying to counter condition the horse to us as a species or individual then training that is measured in seconds and not minutes is the key.
As is training in which the horse feels she can leave if she wants to.
And training in which the food value isn’t so high that the horse over-faces himself because he wants the food.
And training in which the horse feels protected from us and in which we protect ourselves and our horses from our own reflexes or patterns of correction or chastisement or defensiveness.
I’d been a poisoned cue for years with my horse when he met Helen Spence.
It’s normal! Sometimes we behave in ways the horse finds aversive, sometimes we don’t.
Sometimes we bring food. Sometimes when we bring food we also drive the horse away from it and cause the horse to fear us when there is food around.
Sometimes I was asking for behaviour with aversives and clicking and treating for it.
Sometimes I was using target training alone or free shaping or luring.
At other times I was using negative reinforcement and punishment.
But when I saw the response of my horse to Helen, everything changed for me.
He’d never met her before and when he did the experience with her was appetitive from start to finish (apart from one part when he went totally over threshold due to some other nearby horses running around) and in which she and I were in the training area with him. I am sure he did not associate that with us.
But what I saw was his expression when he was with Helen. He was all “You seem really nice. What are we going to be doing?”
With me I could see he was always thinking “Not sure about you. You’ve been pretty horrible to me in the past. How do I know you’ve really changed? But yeah, I’ll touch your hand for some food if I must.”
And to make it worse I had to “Parelli” him back on the trailer to go home after that clinic. I promised I’d never take him to another away from home clinic and I made some other promises to him then as well. Not that it mattered to him what I said. We can promise and praise and apologise with words all we like – but it’s only ever our behaviour that counts.
Watching Archie with Helen was like watching a good friend with a new partner who adores her – and her him, after seeing how different she looks to all the times you’ve seen her when she has taken back (for the umpteenth time to your bewilderment) a husband or partner who has been abusive or unfaithful or selfish. It’s difficult to see how there will ever really be trust in the relationship. The wife is always guarded or expecting the worst nightmare to repeat itself. Expecting to be betrayed or hurt or to be overlooked or taken for granted.
And I felt like it might be for that abusive partner who knows what he has done, and has made all his pathetic excuses for it, but is now watching his wife with her new lover.
It was when all my knowledge of poisoned cues came together. I knew all about poisoned cues but I hadn’t really been looking hard enough at myself as part of the cue.
It was when I saw my horse with someone with whom he had only good associations – probably for the first time in his life – that I saw the way in which we become poisoned ourselves as individuals – or by association – as a species.
Because the way he looked at and behaved around me was very different to the way he looked at Helen.
And I stood there trying hard not to cry, being angry at myself for all the claptrap I’d fallen for in the past and being mad as hell at the people who were perpetuating it. And feeling like I needed to rescue from it all the horses of my friends still doing it. It was a typical grieving process.
We positive training converts and apostles don’t bang on about avoiding aversives with horses for the sake of it.
We don’t do it because we think people who use aversives are all evil or nasty. I firmly believe that the only reason most people are using aversives with horses is because they don’t know what else to do or because they are themselves in deep emotional conflict that is resolved by their apparent cognitive dissonance.
They love their horse but they are also afraid of their horse. Or they love their horse but they care too much about what other people think of them.
Or they love their horse but what they want the horse to do is just 1 percent more important than how the horse feels about it.
And that will be true of all other aspects of their life. The relationship we have with the horse is where everything about our belief systems, our attitudes, our neuroses, our fears and our desires is laid bare.
We don’t bang on about aversives because we want to shame people out of being aversive with horses. It’s because we’ve experienced the shame of the realisation of how the horse is looking at us.
We do it because many of us have been through that grief and felt intense shame ourselves deep down inside. And because the only way out of that feeling is through it. As fast as possible.
The best way to make peace with yourself is honesty. If I could save anyone from the experience of seeing their adored horse fall head over heels with someone else then I would. It’s a killer. It breaks your heart.
But it is also intensely empowering if you choose to let it be so.
I saw my horse fall in love with Helen and it made me determined to BE his Helen Spence.
I saw that look and I said to myself “I want THAT!”
We’re a work in progress, but I think we’re doing OK. Actually no, I will correct myself. We’re doing really great.
The Effects of Combining Positive and Negative Reinforcement during Training – Nicole A. Murrey https://pdfs.semanticscholar.org/26ba/1a8d8ee2c088af43f80e10b7e0f65748cd01.pdf
How the choice of reinforcement effects the perceptions horses have of humans. Sankey, C., Richard-Yris, MA., Henry, S. et al. Animal Cognition (2010)
Pingback: Poisoning – or counter conditioning? – Illis ABC