What now?

Predictability, all the things that tell horses what to expect, and what to do next. When we’re training them they rely on the clues, such as where they are, where any apparatus is, what their person is doing, what other persons (including horses) are doing, and so on. Discriminating exactly which stimulus (clue!) reliably accompanies a behaviour and its outcome is a process of elimination. The horse has to learn a stimulus-behaviour-outcome sequence and to do so he has to turn detective.



Can horses really be that bright? Mischka, a 22 month old sports horse filly is learning to touch cones with her muzzle for food, then to learn to touch them in alternation in order to earn the food, and then to perform the same pattern of alternation with an increasing distance between the cones.



That’s hard. At 22 months of age Mischka is a horsey teenager, puzzles like this can easily frustrate her and make her either quit the training process or become enraged by it. She really needs clues as to what to try next. The first clue is that her person is looking at one of the cones which are placed side by side; their position will become another clue. Mischka investigates, her person says YES! and immediately presents her with a handful of food. Miscka investigates the same cone again but her person is now looking at the other cone. Miscka investigates more thoroughly in case her person hasn’t noticed then switches to investigating the same cone as her person. This makes her person say YES! again, and more food is presented. The person’s attention returns to the first cone and after a trial investigation of the second cone, Mischka returns to investigate the first, with another YES! and more food. The clues Mischka used to get the right answer were the direction of her person’s attention, the position of the cones, her preceding behaviour, and the behaviour reinforced via the reception of food.



Mischka rapidly makes the link between alternating her touches of the cones and food, or no food if she makes the incorrect choice, and consistently swaps cones. The stakes are then raised. The cones are placed nearly half a metre apart. Now Mischka has to make an effort to get her nose from one cone to the other. Perhaps it’s easier to stick to just one cone now? Mischka tries this and only touches the first cone after being rewarded with food the first time. There is no outcome for this behaviour and Mischka makes the effort to bring her nose to the second cone. This makes her person say YES! and food is delivered. Mischka recognises the clues from the earlier lesson, notices where her person’s attention is, recognises that touching each cone is only rewarded on the first touch, never the second, third or any other number of touches, then smoothly returns to alternating her touches.



This raise in the stakes is repeated several more times, with the distance between the cones growing to nearly 8m. Now it becomes very apparent that one of the clues that Mischka is using is the proximity of the other cone. When it’s further away and slightly out of view because it’s behind her, the behaviour of switching between the cones falls apart. Swinging her forehand from one cone to the other no longer works because the stretch is too far and it puts the previous cone out of sight so that it is no longer a clue that it is also the next cone.



Mischka first tries to solve the problem by walking to the people in the training area, perhaps to get a better idea of what they are looking at! Her attention seeking behaviour is ignored and then Mischka spies the other cone. She’s told YES! just as she starts to head towards it. This gives her the clue that she’s got the right answer, but she continues her route and touches the cone before looking for her handful of feed. The clues to keep going are the early signal for reward, and her person looking at the cone and also walking towards it where she ultimately gives Mischka the food. Mischka tests investigating the people again when she finds them closer to her than the next cone, and after a few trials gives up as there is no outcome.



The sight of the next cone remains an important clue. On one occasion Mischka starts to walk to her person after leaving a cone, then remembers that this will not result in food, so stops and starts to walk backwards. While reversing she notices the next cone and turns and walks forwards to touch it – with the now firmly expected outcome of food reinforcement. After consuming the food Mischka walks backwards again, because last time she got reinforced for touching a cone she had just happened to be walking backwards in the same place. Mid back-up she steps forwards again and touches the cone she just backed up from. Her person fails to reinforce her because it was the incorrect cone. Mischka walks forwards then and circles around finding the next cone, and some positive reinforcement for doing so. Then she reverses, catches sight of the next cone and walks onto it.



Reversing has become a means of sighting the next cone, and a clue that she is going to make a correct choice. So also does circling forwards and away from a preceding cone. Clearly a less changeable ‘go find the next cone’ signal is required. Mischka needs to discern just one signal or specific compound of signals that means ‘there is another cone for you to find and touch’. Her person adds in a verbal signal ‘TOUCH’ while focusing on the next cone just as Mischka is on route to it and success, regardless of whether she reversed to sight the cone or simply walked forwards from the preceding one.



The next chapter is where Mischka learns to look for the cone in direct response to her person’s direction of focus and the verbal cue, perhaps even if the next cone has been moved a little since she was last there.