In these articles, I describe the ways in which horses learn and do not learn, and how learning can be hastened or impeded. In behavioural psychology, learning is defined as a more or less permanent change in behaviour following reinforcement. By reinforcement, I mean something akin to a reward or punishment which channels learning in a particular direction. The equine brain is marvellously evolved to learn different sorts of information in different ways. As I have described previously, the horse has a fantastic 'photographic' memory of the physical world, and also of his responses to it, which manifest themselves as rapid-fire habits, or automatic responses. This memory is maintained through quick access to long term storage, and is unencumbered by the clouding effects of reasoning and imagination, keeping reactions precise and ever ready.
The expression of these reactions is effected by instinctive behaviour
such as the flight response (fear), the social affiliative drive (the
horses desire for companionship), the hierarchical drive (pecking order),
the sex drive, hunger, thirst and many internal aspects such as health,
robustness, hormones and so on. When it comes to teaching the horse movements under saddle there is one major stumbling block which riders are inclined to ignore, and that is their own balance. Nothing but plenty of practice in correct position, will teach you the finer points of exact balance. It's not advised to learn balance on a young horse, as he will also learn bad habits such as crookedness from your shifting centre or balance, and forward problems and a hard mouth from your attempts to keep your balance with the reins. Precise balance in the upright and forward seat must be constantly monitored by someone qualified to do the job, and the short-cuts lie only in teacher effectiveness. Only when the rider is in balance can he perceive when the horse is in balance and correctly reinforce the horse when he too is in balance.
The outcome of learning is in the establishment of habits, or automatic responses. Habits exist in the brain as nervous pathways and specific chemical substances. In the early stages of habit formation, these structures are fragile, and become more solid and permanent with repetition over time. In general, it seems to take around three to seven consecutive repetitions to begin the process, and more to see them begin to become reliable, provided that the horse is adequately and immediately reinforced, which may be in the form of freedom from pressure (softening of the hand and leg), food or other primary reinforcement such as scratching at the base of the mane.
The stronger the reinforcement, the faster the setting of the habit. It is important to avoid the fear response at all times for fear is such a strong reinforcer that habits associated with it set quickly. In training, the formation of habits can be hastened by adopting training schedules of three to seven repetitions of one task, then maybe a walk for a few minutes on a long rein, then back to the previous seven repetitions and so on until you have achieved at least three sets of seven intervals for each task.
For instance, if the horse refuses to strike off on the correct leading leg at the canter, (a common ex-racehorse problem), then the idea is to teach him to take the correct lead through say, shoulder-in, or maybe through lungeing or by cantering over a pole on a circle to the right, (keeping the neck straight and maintaining supporting outside aids), then repeat it in the repetitious schedule mentioned above.
Such a problem ideally should be solved with the help of a qualified coach, for the problem is generally exacerbated by crookedness and imbalance. In general, the cooler and less reactive the horse's temperament, the slightly greater the number of repetitions that are required for any given task. Especially in behaviour modification, I have found learning to accelerate significantly using this strategy.
Another important aspect of repetition is to follow on from one repetition to the next almost immediately. The canter strike-off, for example, will be more successful if the next repetition is attempted while the pattern is 'still in the horse's head' from the previous repetition. The longer the delay, the less likely will the correct new response occur.
To give another example, the horse which has refused the jump (provided he is not overfaced or impeded by the rider) should be made to retake the jump as soon as possible followed by at least three immediately rewarded repetitions, for learning to be accelerated, so that the tendency to stop will be extinguished. The longer the delay in between repetitions, the less the horse learns from the correction, and the more the horse will learn that stopping is available to him. I believe teaching the horse to be willing in this way is an important attribute for an event horse.
To eradicate habits, the correct approach is to prevent their expression. For example, if the horse habitually 'falls out' through his shoulder at a particular point on the arena, the rider should prevent it happening by applying stronger counter aids just before that point next time, then soften and reward straight away. These measures should be taken for the next three times at least, just to be sure that the tendency is not still there and that the habit is not beginning to form. It is no use trying to deal with the problem after the point where the habit has been expressed.
In all cases, when correcting habits in horses, care should be taken to find the root cause of the habit, for then, the habit, once eradicated is unlikely to re-occur.
Rearing is a habit generated by the horse not going forward, and most frequently by not being allowed to go forward by harsh hands. Once learned the horse discovers a clever evasion and actively uses it to avoid going places. Rearing also occurs in some horses which during their breaking-in were not taught to go forward with big active steps in the walk in the first place, setting up the conditions for problems later when the horse is put under more pressure to go forward by the rider.
Reinforcement is a word used extensively in learning theory, and as explained earlier, it basically means something which increases or decreases the likelihood of the associated behaviour occurring again. As Tom Roberts so accurately puts it:
" If the horse 'profits' by receiving food or companionship, or by the cessation of work or pressure, or by the gaining of freedom in the form of escape, if he even delays exertion, then his associated behaviour is reinforced, and therefore likely to be repeated." For example feeding the horse after he comes to you increases the likelihood of his coming again to be caught. Reinforcement is divided into primary and secondary, and positive and negative.
A primary reinforcement is one which acts directly on the instincts of the horse, such as food, touch, pain, whereas secondary reinforcements are those which are associated with primary ones through learning, eg. patting, voice praise such as 'good boy', or 'No'. Secondary reinforcements have no power in themselves unless the horse has learned to associate them with the primary reinforcement. This is why patting horses is a humanistic and inappropriate way of rewarding behaviour compared to scratching the base of the mane, which was proved in a recent study to produce relaxation in the horse more than any other area.. In training under saddle, food rewards are not always practical in that they can lead to the expectation of food and the consequent loss of carriage or posture leading to further unwarranted behaviours being incorporated into the trained repertoire. So scratching in front of the saddle is the reward panel right at your fingertips.
Giving the horse something it likes during or just after desired behaviour is called a positive reinforcement. It is distinguished from reward by the exactness of its timing. 'Reward' is a sloppy term, for it does not imply any timing necessity (people reward their horses milliseconds or many minutes after the act!)
Delayed reward is not effective compared to positive reinforcement because of the difference in timing, as the horse simply cannot make the connection. The only useful outcome of delayed reward is in the effect of reinforcing your bonds with the horse - which, in itself, is perfectly desirable. But don't fall into the trap of thinking that the horse will learn from a delayed reward: titbits after a good workout strengthens bonds between you, but do not serve as a reward for the workout.
The best rewards for good work are those given during the work, precisely after the desired response. A loose rein and a good scratch within a second or two of the correct response are the most powerful rewards available to the rider, that is unless you can dismount within that time frame, as getting off the horse's back is also rewarding!
Most of the horse's training is accomplished through negative reinforcement, but do not confuse it with punishment; it has nothing to do with punishment or violence. Negative reinforcement is defined as removing something the horse doesn't like (eg your leg pressure) to produce the desired response (eg leg yield), the moment the horse complies.
Negative reinforcement occurs during the undesirable behaviour and not
after it. It may be the tap-tap-tapping of the dressage whip which is
used as the negative reinforcement, ceasing exactly the moment the incorrect
response finishes and the correct response begins. By contrast, punishment is the delivery of a painful stimulus after an act. The use of untimed punishment in horse training is in part a reflection of the retributive mind set of our cultural tradition, and is based on the unfair and false premise that the horse can connect cause and effect relationships of events separated by time and so when you make him suffer for his misdoings the connection will be made with the behaviour which preceded it. In fact, punishment is only effective if it occurs immediately after the incorrect response, and if it is used judiciously and once only. It should not be used for anything other than dangerous resistances.
The outcome of delayed punishment is that the correct response will not be incorporated into memory, and the fear response will be on red alert, depending on the degree of pain inflicted and fear will be sequenced in that particular response. Punishment serves largely to make the punisher feel better, as studies show that for some humans, violence can be reinforcing.
To see the difference between negative reinforcement and punishment, take the example of the horse that kicks habitually when you touch near his hindquarters. Punishment (ie after the act) will only make the horse more fearful and adrenalised, he will likely kick harder, and the lesson might go on for ever.
Consider the approach, however, of negative reinforcement: just tap a long dressage whip on the horse's flanks and progressively down his legs during the kick, removing it only when he stops kicking, and the kick will reduce to nothing.
In behaviour modification, behaviour mostly tends to reduce in intensity, rather than disappear immediately. So bucks or rears become reduced to token gestures, then finally disappear. This is because the behaviour is generated by associated habit, and neural pathways don't just disappear, they take time to be diverted, like gradually changing the course of a river by diverting it.
In correcting problems, this gives us some lead time to teach the rider how to maintain the new behaviour, and to be taught by the trainer to identify the remnant signs of the old vice and nip them in the bud. Behaviour, both good and bad, is generated by the acquired collective presence of certain brain 'wiring' patterns, which were placed there by the physical world acting upon the horse, most commonly they are put there by the rider.
You must be sure to reward the horse's every attempt at desirable behaviour by the power of touch coupled with the words 'good boy' etc. Despite his brilliant memory, his limited powers of connection mean that you should not blame him for his behaviour. 'Decisions' to act in certain ways by the horse are not made by a reflective approach as they are in humans, but instead are a direct representation of the most reinforced neural pathways in the brain of the horse, in concert with his natural drives and instincts such as fear, dominance, curiosity drive and so on.
Appreciating this will go a long way in understanding the induction of good habits and the extinction of bad ones, for when viewed in this light, the horse is not responsible for his behaviour through free choice, but instead is very much a victim of his learned responses.