Negative Reinforcement (pressure and release) training works, no one is denying that.
But we have to be honest with ourselves and put ourselves in our horse’s hooves.
If someone squeezes our arm to make us move out of their way and if that squeezing (and proximity) causes us to move and they promptly stop squeezing AND we are likely to move away in the future, that’s Negative Reinforcement. Most importantly, it’s not like putting a coat on when it’s cold or drinking water when we are thirsty or choosing to go dancing with a loved one. This is someone else making a decision for us, about where we should be standing, by using some kind of physical, emotional and mental “pressure” (unease/discomfort).
Imagine that same person does all kinds of things like that to you, after a while they don’t even have to touch you, they simply reach for you and you move away. You really don’t want them squeezing your arm every time. You might also get a little twitchy and hyper vigilant when they are around. This is so you are ready to do something, before they have to touch you or pressure you or make you do something you probably didn’t want to do or like to do. We call that avoidance behaviour.
Now imagine the arm squeezer (plus various other methods of discomfort to make you do various other behaviours) keeps you in an enclosed area, away from family and friends, dictates when you leave the paddock, when and what you eat, who your friends are if you have any, what else? When you need medical treatment, if they are even aware you need it, because they think posters about equine stress and pain are just dumb. Their horse does it all the time, so it can’t be stress or fear or pain! Or worse, they tell everyone it’s good and great when they get a calming, appeasement or stress behaviour, because the horse is “communicating” and then they promptly ignore what the horse is saying and halter them and get on with their pressure and relief training!
Imagine a world where we said to children, we’re not going to force your beautiful pony with kicks and crops like those other people!! We are going to train them in a way that you BOTH have fun, by using Positive Reinforcement training, with clicker training, by training with food and scratches!
Imagine if we said to children, you’ll have more fun on the ground and get to know your pony better that way and build a friendship. You’ll build a real relationship, where your pony will truly love you back.
Imagine if we taught them how to have fun and laugh with their pony and the pony was laughing and having fun too.
Imagine children who learnt empathy for all creatures, not just how to use them and make them do what the child and adults want.
Imagine a generation of children growing into adults, who could ‘talk’ to their animals, by understanding how all creatures learn and that they feel emotions too.
Imagine children who didn’t have to cry in frustration or in pain, from falling off or being bucked off unwilling or scared ponies. Imagine a miserable and scared child, not having fun and not wanting to do things with their pony and being told to push through the fear and pain?
Instead, imagine children learning via Positive Reinforcement (clicker) training, how to build important life skills as humans, such as empathy, observation, problem solving, analysis, hand eye coordination, thinking on their feet, emotional regulation, developing reciprocal relationships and positive interactions with others.
With all that learning, we could set them up for life with amazing skills that build their confidence, self esteem and physical and mental wellbeing and their emotional maturity and regulation. As well as learning how to interact with others in a positive way, where both parties can ‘win’.
Imagine us being the adult heaping Positive Reinforcement on our children! Saying “well done!”, “you were awesome!”, “you worked that out so well!”, “you’ve made your pony so happy and well trained!” and “I’m so proud of you!”. WE adults enjoy the benefits as well, because it makes us feel good too!
Where do we get so lost, that we insist children hit and kick their pony for their own pleasure, who then grow into adults who do the same?
Worse, the child turned adult, who mostly doesn’t enjoy what they are doing with their horse, feels mounting pressures from the equestrian world and discontent with what they’re doing. Therefore they force the horse more and more, who feels the same.
When we clicker train, there’s lots of different ways that we can convert what we are doing on the ground, to ridden behaviours. I thought this video was a great example of true liberty work. No whips directing her shoulders or body, no need for extensions of my arm, no body pressure and no “do it or else!” cues. We’re having a conversation and we’re both a bit rusty at the start.
In the old days before clicker training, I used to do a lot of riding with Mercedes where she moved around based on where I was looking and where I turned my head, shoulders, hips and that influenced my weight distribution in the saddle, rather than using pressure rein cues. If I looked around and moved my shoulders, my hips followed and there was a subtle shift in the saddle from my seat and legs. Mercedes would turn and circle on these cues.
In this video, you can see the same cues on the ground, all trained with Positive Reinforcement and without any aversive pressure, whips or other tools. I am usually much more subtle than this, but we haven’t done it in a while. I’m ensuring she is successful and the most important fact is that these are appetitive cues. The consequence of her performing the behaviour means food is coming and if she doesn’t get it quite right, food still comes and I work on refining my cue.
When we transfer behaviours and their cues from the ground to under saddle, there’s always a bit of a transition stage, where we might need to use a target stick or exaggerate rein cues a little, or have a helper on the ground, until the horse ‘gets it’. But from there, we can refine the cues to whatever we want and for Mercedes, she remembers the feel of my head, shoulders and hips moving to indicate the direction and I can transfer big obvious cues for things like turns, such as an opening rein and transfer the cues back to what my body is doing. But this time they’re appetitive cues, because she gets food when she does the behaviour after the cue.
Watch the video and imagine me doing what I’m doing on the ground next to Mercedes, and lift me up and put me on her back and that’s what I do under saddle and Mercedes does the same turns and flexing and bending. She’s super smart and because of all the clicker training we’ve done, the old body cues no longer cause her any worry. Much like the mounting block video I took, all the old unhappiness from past training has disappeared and been replaced by all the enjoyment of the clicker training with food.
There’s also a lovely synchronicity in how our heads, shoulders, hips and feet match up.
Something else to notice as well and why taking video is so important. I saw things in my video that I kind of noticed at the time, but seeing Mercedes behaviour now, I am continually blown away with how clever and subtle she is!
Let me critique myself!
I can see some poor feeding technique, where I’m feeding out of balance (with the wrong hand), feeding too close to my body and there’s a few misfires because we simply haven’t done a lot of this type of training lately. I blame my back problems on poor feeding technique and the rest is simply lack of practice and doing lots of other training that’s been very different to what we are doing here. Oh, and it started raining!
What I like in this video, is that at one point early on, I noticed Mercedes wasn’t standing very straight and I glance at her feet. It was like she read my mind, although I know it was that mere glance I gave her feet, that prompted her to step across and straighten up slightly. It’s something we’ve worked on in the past and it was very cool that she pulled that out of her repertoire with very little thought or prompting on my part.
Then another time when I switched sides, I realised there was some latency, she was slow to respond and not sure what to do. We got stuck at one point and if we get stuck, I’ll give her a handful of food to keep her feeling successful and to prevent frustration and see if she can figure it out. What happened was she offered a small movement of the head and even though I didn’t click for it, I fed her and that was enough to tell her she was on the right track and she took a step to her left and we got moving again. I’m sure I only fed her to keep her ‘in the game’, but watching the video, it seems like I confirmed her idea that we were supposed to be turning left. It’s always good to practice equally on both sides and I’d say I’d been favouring one side too much, based on her response.
To recap, horses are all about their environment and where their body is in their environment. That means that they will see and feel subtle cues. Simply our shoulders turning and our body following either next to them or on top of them is going to elicit a response. We can use our body to communicate and it doesn’t have to be a threat, if the consequence of behaving is food, pure and simple.
FF R+ dog trainers talk a lot about encouraging people to comfort their dog when they’re afraid. Patting and praising them when they’re scared, wont reinforce their fear.
I’m a FF R+ equine trainer and I say the same about horses. If your horse is afraid, giving them food (or patting and praising them), won’t reinforce their fear.
Similarly, there seems to be a lot of equine clicker trainers that are very much focused on trying to reinforce “calm” when they start out.
I’ll admit though, it does sound kind of convincing.
But as long as we focus on trying to train or reinforce something that the horse feels, by withholding food for that elusive emotion, we stray further away from what we should be focusing on instead.
There are many trainers who are waiting and watching for “calm”. They are withholding food because they are waiting for the emotion that they can supposedly reinforce.
It’s not a good idea because we can’t reinforce an emotion, but it’s not telling the horse what *to do* either.
Waiting and withholding is never a good idea. Worse, it causes the horse to try to figure out what they have to do to get the food. It causes the opposite effect. We end up with unwanted behaviours and often with the opposite emotion to what we actually want. We can cause the horse to feel confused, frustrated and tense – the opposite of calm.
What I suggest is always being generous with food, be prompt and catch them doing good stuff and reinforce those behaviours, before the horse starts improvising and throwing behaviours or worse, starting to try to sniff, nudge and take the food from us.
When we train in a way that they’re successful, there’s plenty of food and there’s clarity about what they’re supposed to be doing, then you’ll get calmness.
Remember, we reinforce behaviour and then we can create happy emotions through good generous effective training.
But do we truly understand them and what their needs are?
One of the most important requirements that is often misunderstood is that horses are complex social herd animals. They would never choose to live alone, they need company! In a herd, they enjoy a rich social life, such as playing, grooming each other and even interacting with horses they may not like. This enables them to develop social skills such as reading body language, learning to be a horse and getting along with other horses.
Herd living provides lots of enriching opportunities. Smelling other horses, their manure, tasting and mouthing each other in play and mutual grooming provides lots of opportunities for enrichment and also for exercise, which is often lacking in domestic situations with lone horses.
Herd living is also important not only for direct body contact, but in order for horses to feel safe grazing or sleeping and getting their REM sleep while others keep watch. A lone horse in the wild would be easy prey and would not last long, so stress levels can be high for horses living alone – they can never fully relax.
I realise that we can only do the best we can, I have horses who don’t necessarily get along all the time or given a choice, would not choose to be together. BUT Positive Reinforcement training can achieve a lot towards helping animals live in harmony.
Classical and Operant Conditioning can be utilised to change emotional responses to other animals and help them get along and enjoy each other’s company. If we can train voluntary blood draws and other unpleasant experiences, I’m sure with a little imagination and lots of planning, we could arrange the environment to teach animals to get along with each other. I taught my horses to like my donkeys and they were terrified of them at first sight!
It’s truly wonderful to watch horses and even more wonderful to see them enjoy each other’s company and just be horses!
One thing that I really love about R+ training is shaping behaviour and putting it on cue and playing around with cues.
I’d encourage all of you to develop your shaping skills, allowing the horse to offer behaviour and most importantly, get those behaviours on cue.
All my horses and donkeys were trained (shaped) to walk on a cue, no gear, all at liberty, no targets and grass is no problem either and then I match their footsteps. In this way, it is a built in start button, because if they don’t move off or are slow to do so, I know there’s a problem. If I’m ahead of them and luring them with the food, I take away choice and autonomy and I also lose valuable information if they are slow to react. The time between a cue and the behaviour being performed is called latency. The slower they are to do the behaviour after the cue, the higher the latency, which is really valuable information to us. What is preventing or punishing them performing the behaviour? ie. high latency.
Can we just stand there and cue our horse to walk or run?
If we can definitely say that we simply stand there and cue our horse to walk and they walk off, we know it was trained and on cue. Our horse understands the behaviour and the consequence is food at the end.
Predictability and understanding how to gain the food (what is the contingency) is what makes it fun. Not knowing and having to follow, to chase and even run, is not fun.
Training and clear contingencies for the horse ie. when I do this, that happens (food is given) is what prevents frustration and confusion in the horse.
But if we start walking or running (away from the horse) and the horse follows or chases us, do they understand the behaviour? Or are they just chasing the food that’s moving away from them? Are we just luring them with food?
It’s also important to look at the body language of the horse. Do they look soft in their face? Move with decisiveness and focus because they understand how to gain the positive reinforcement? Does their whole body look soft and relaxed? Do they nicker with enjoyment? Lick their lips in anticipation of the food, not out of tension?
Or do they flick their head, pin their ears? Does their whole face and body look tense? Do they trot, canter, rear and even buck? Do they look like they’re truly having fun?
It’s easy to walk around with a food pouch and your horse follows you around. It takes a lot of good observation, knowledge, timing and skill to shape behaviour without discomfort to the horse and then put it on cue.
This video is a good example of how I have previously shaped the behaviour of walking and trotting using the reverse round pen and now I am then adding the cue for the trot. It would be easy to think she is trotting because I am giving the cue, but that’s not what is happening. When the trot behaviour is offered consistently, then I add the cue. Timing is super important. I need to say the cue the moment she looks like she is committed to trotting, so she makes the association. If I say it too soon, it won’t have relevance and could actually interrupt her wanting to trot and saying it too late won’t be relevant either. The verbal cue is the word “faster” drawn out a little to give her time to fully transition into the trot.
I often see videos and descriptions of happy horses, or playing horses or horses generally having fun.
But when I look at the horse’s body language and the context, where they are often in an arena with a human, a human who is often waving a whip or their arms around and running about, sometimes even wearing a food pouch. Or it might be a horse alone in a paddock or arena, let loose in a strange or scary environment, then I don’t think it’s fun for them at all. Often there’s a lot of conflict behaviours being shown.
Horses can play, but it’s with other horses and often it’s static play with lots of attempts to bite and head jousting alternating with quick bursts of chasing and then more static play *with other horses*. Good fun play is always reciprocal, they take turns.
Head flicks, arched necks, tail swishes, tense faces, walking away, staring into the distance and chasing humans with food, is not evidence of fun or play.
Always question and always look at the environment and context of the behaviour. Sometimes putting yourself in the horse’s hooves can help, even though we are not horses, but we are both social emotional beings. We don’t like to be teased or chased or pressured or threatened either.
I heard an interesting statement that I want to tease out a little, because I’ve heard variations of this statement many times, particularly from those who are against training with Positive Reinforcement, without pressure or force.
Giving a R+ trained cue and then most importantly, HOW they respond, gives us information. A cue is not a command, it’s not a “do it or else” and it’s not an ultimatum without choice. A cue is a polite ask and it’s their choice if they want to oblige. A cue in R+ training *ideally* has choice built in.
The amount of times I’ve started off posts saying, “what I love about Positive Reinforcement training” is a bit embarrassing.
But there is so much depth and nuance to the training and sometimes it makes me sad that people are so dismissive of something that is so complex, nuanced and organic. It’s also extremely empowering and pleasurable for the animal being trained and it develops a communication between species that no other training can replicate.
What I also like is that you don’t have to be an expert ethologist or reader of body language to train with R+ and know what your animal is telling you in a training session. All behaviour ie. responses that have been observed in an R+ training context, have been analysed and have had a name given to them that we can discuss. It’s not new or magical to observe an animal’s behaviour and identify how they are responding and interpreting our behaviour and our cues in R+ training. Latency is a good example of this.
What I find interesting is that detractors of R+ training claim it is manipulative, coercive, controls the animal and their mind and that there’s no choice. Do it for the cookie or else! R+ training is quite the opposite actually, if done well.
A behaviour trained with R+ and put on cue is not an ultimatum. In the hands of a good trainer, it is anything but.
There is a famous quote attributed to Viktor Frankl which I have since learnt, was a quote describing some of his work, but not actually directly quoted from him, “Between stimulus and response there is a space. In that space is our power to choose our response. In our response lies our growth and our freedom.”. What I like about this quote is that I relate it to choice in R+ training. If the ‘stimulus’ is the cue for the behaviour (Discriminative Stimulus (SD) ) then the space refers to how the animal *chooses* to respond to the SD/cue at that moment in time. HOW they respond, gives us information. A cue is not a command, it’s not a “do it or else” and it’s not a given. A cue is a polite ask and it’s their choice if they want to oblige. This is the opposite of pressure trained cues, worse, the command or demand is the way that the behaviour is elicited as well. There’s not much choice in that – discomfort or what?
But when we train a behaviour with R+ and put it on cue, if we’ve trained it in a way that met the animal’s needs, provided alternative sources of reinforcement, be it positive and possibly negative as well and watched and listened to their behaviour and body language and responded appropriately, then we’ve set the stage correctly. The most important thing in my mind is responding to their behaviour and body language, because that’s where the communication starts and where we can tell them that we are listening, we are responding and we are offering them alternatives and choices.
The other important thing in providing choice in training is to approach the R+ training with an ‘errorless mindset’. The animal is never wrong, what did we do? When we avoid withholding positive reinforcement (food) for “incorrect responses” and look at all behaviour as communication and that we want to encourage and even reinforce that communication, then we not only open the doors of communication, we tell our animal they have choices.
In practical terms, if I’ve trained a behaviour and I know it’s on a reliable cue and my animal does not do the behaviour when cued, or is slow to do it, or offers a different behaviour or sadly, shuts down, then that’s their choice in that moment, in response to the stimulus. What I do is give them food anyway and then either cue a super easy behaviour, click and feed and end the session, or end the session straight away (possible Negative Reinforcement or Negative Punishment) leaving a generous amount of end of session food, or I might cue the behaviour one more time, in case they just didn’t hear, see or recognise it. But I’ll only cue it one more time. Any more than that and it can turn into pressure or a command or a nag, and that’s not fun. I’ve already done a post on this, focussed on discussing errorless learning.
Then it’s up to me to figure out why the animal didn’t perform the cued behaviour. I often find there is something urgent and important happening for them that over-rides the cue, something like pain. This is another thing that detractors like to say, but good R+ training means we don’t train away fear and we don’t ignore pain. High latency, lack of alacricity in performing behaviours, performing other behaviours or even calming signals, are all signs there is a problem and we don’t keep giving the cue and ignoring what the animal is saying to us. This is where choice comes in.
What happens when the animal has choices in training is that we give them a level of control, IF we listen to them and their behaviour. Choosing to respond in a certain way is controlling their environment.
Helping a horse who is experiencing fear should never be about making them “face their fear” or “get over it”. What can fear look like? It can look like anything from a horse freezing, refusing to go forward, baulking, napping, looking off into the distance, leaning away, rubbing their face on their knee, licking and chewing, yawning, looking generally tense and distracted or trying to run away. Any kind of attempt escape or avoid an object, location, situation etc is a good sign the equine is not comfortable.
The first thing we need to do to help horses experiencing fear, is to rule out pain as a reason. Pain can cause exaggerated or unexplained fear and/or aggression in animals, so it’s a good idea to eliminate it as a reason first.
Then we need to take a systematic approach. This means introducing the fear inducing stimulus at a distance/exposure that the horse only just notices it, no closer or stronger. At that moment, we pair the exposure with food. Incrementally we decrease the distance/ increase the exposure, so that in time after many pairings with food, the stimulus comes to be the predictor of food. This is called Systematic Desensitisation and Counter Conditioning.
We are systematically desensitising them without the horse even realising, keeping them well below their fear threshold at all times. At the same time we are turning their emotional response into a positive emotional valance. The horse sees the plastic tarpaulin for example, as a predictor that good things like food are coming soon and therefore causes a rush of pleasure.
That’s got to be better than being made to “face your fear”!!
One thing that’s very important is that we DON’T try to get the horse to go closer or *do something* for the food. For example, with the tarpaulin, we wouldn’t try to lure them towards it with a nose target and they only get the food when they step towards it or touch it. Worse, don’t scatter food on the tarpaulin and make them “face their fear” in order to get the food. It’s always best to do a simple pairing of food, not make the horse have to do something around something scary, before they can have the food.
Further reading on Desensitisation and Counter Conditioning with thanks and credit to Eileen Anderson at eileenanddogs.com :-
Although people struggle with the terminology of R+ training, I’m a big fan of using correct terms, not for the sake of it, or to look super brainy, but because it improves our training. Knowing what something is, or what to avoid or the name of something you want to use, is important in being a better R+ trainer.
I encourage everyone to learn the terminology.
The term and skill I want to talk about is Contiguity (noun) / Contiguous (adverb).
It’s important to note that Contiguity occurs in Classical Conditioning and Operant Conditioning and as usual, Pavlov is on one shoulder and Skinner is on the other. That means that we can talk about how it occurs in Classical or Operant Conditioning separately, but both things are happening simultaneously.
In simple terms, Contiguity is the space or time that elapses between one stimulus and another.
A simple example of this is when we pair the sound of a click (one stimulus) with food (the other stimulus). If there is very little time that has elapsed between the click and food and we do it repeatedly, click – food, click – food, click – food, the faster the learning and the faster the association is made. This is a pairing procedure (Classical Conditioning) and the time between the sound of the click being perceived by the learner and the moment they acquire the food is the contiguity. There have been studies done on the ideal window of time that has elapsed between one stimulus and another, for optimum learning, but I would like to just focus on the concept and ideally we are talking about adding the second stimulus as soon as possible after the first. What then happens is that the first stimulus, after repeated pairings with the second stimulus, comes to represent and conditions the same response as the stimulus that follows. Imagine after pairing/conditioning, that a horse salivates at the sound of the click, before the food is even delivered, much like Pavlov’s dogs salivated at the sound of the tone, before the meat powder was given. Operantly, the sound of the click becomes a secondary reinforcer due to the pairing/ association made (Pavlov AND Skinner happening).
An Operant Conditioning example would be the time that elapses between Behaviour and Consequence.
I’ve talked about Latency in a previous post, which is the time that has elapsed between the Antecedent or Cue (Discriminative Stimulus) and the Behaviour. Contiguity is about the time or space between Behaviour and Consequence.
The first time I learnt about using contiguity in clicker training was when I did my very first online course with Peggy Hogan to learn about Shaping and one of the exercises involved not using a clicker, but simply delivering the food as quickly as possible. I don’t recall if Peggy used the actual word, but I do remember when I did Susan Friedman’s LLA course, that the light bulb went off when she discussed Contiguity and I knew that’s what Peggy was teaching me.
You can see the video where I recorded the fun exercise Peggy had us do, where we pre-loaded our food (a whole other subject I might cover later) in order to deliver it as quickly as possible after the horse did the behaviour. In this case, the Criteria was any interaction with the chosen object, in our case, it was a box. So many times people ask me when they watch the video, are you clicking? I’m not, because I don’t need to, thank you Contiguity! That’s the beauty of clicker training really, we don’t always need a clicker and there is more than one reason why this is the case.
There is also a number of other concepts being learnt and illustrated in the video aside from pre-loading the food and Contiguity. There is a number of other things happening and being learnt as well, one of the obvious ones is called “Feeding for Position” as Bob Bailey liked to describe it.
Another thing that’s useful in understanding about Contiguity, is when you are clicker training and you realise you missed a moment to click, don’t click anyway. It’s always best to deliver the food as quickly as possible to reinforce the ‘right’ (goal) behaviour or an approximation and not click for potentially the ‘wrong’ behaviour or approximation. If we want our animal to understand the click and we want it to be a powerful and precision instrument, don’t throw it around randomly. Use it carefully and sparingly in order to develop and safeguard its power and precision. We can do this because of Contiguity!!
One of the key reasons Contiguous Reinforcement is so important is because when there is a large enough space or time gap between the Behaviour and the Consequence, it enables the animal to behave in that gap. If you’ve ever reviewed video, you will note that behaviour can change within a second and different behaviours can be performed seconds after the goal behaviour. We need to be quick ie. Contiguous in our Reinforcement, in order to avoid Reinforcing other or unwanted behaviours!
In laboratories, in controlled conditions, delayed Reinforcement can still equate to effective and efficient learning. But out in the field, with so many variables and other potentially reinforcing (or even punishing) stimuli present, the shorter the time/space between Behaviour and Consequence, the better.
This is yet another reason why it’s a really bad idea to withhold food for better or ‘perfect’ behaviour in clicker training. All you risk doing is reinforcing ‘superstitious’, or a chain or medley of behaviours that were performed between the goal behaviour and the delivery of Reinforcement.
Prompt delivery of Reinforcement (food) after any approximation of behaviour, whether you click or not, is always more effective and efficient than holding out for more.
