Horse training traditions & a constructional graded approach emphasizing minimally aversive methods
Horse training is frequently guided by traditional approaches, trial-and-error, and anecdotes, rather than behavioral science. Significantly larger than their human handlers, aversive tactics and training instruments are frequently used to subjugate horses. Naturally a social and cooperative species, despite their size horses are impressively adaptive; they are potentially the fastest learning and most trainable species of domestic animal. While humans and horses have similarities (relying on reciprocating and close relationships for survival), the difference between being a prey versus a predatory species can lead to misunderstandings and miscommunication.
‘Dominance’ and ‘disrespect’ are two words commonly used to describe undesirable behaviors, with many popular trainers dictating that the horse needs to ‘respect’ their handler… However, horses do not have the same concept of ‘respect’ as people do, so this is senseless. Studies of horses in the natural environment find that although certain animals are bossier or more experienced then others, in general the herd makes decisions as a collective group. Horses are also very comfortable being in close contact with each other and use grooming and touch to bond. Problematic behaviors directed towards people, often labelled ‘dominance’, such as striking, biting, and bucking, are frequently side effects to aversive stimulation and connected to escape behaviors.
The 21st century has seen exciting changes in transitioning how horses are trained, with an emphasis on less coercive and more humane training practices that emphasize developing trust between horse and handler. Positive reinforcement is still less common in horse training then seen in dogs and other species, however it is not as dismissed as a training technique as it was previously, although there is still considerable misunderstanding regarding the use of treats in horse training within the equestrian world (concerns over crowding, ‘disrespect’, spoiling, biting, etc.). Used correctly, positive reinforcement has been shown in numerous scientific studies across species to be superior to aversive or mixed methods in speed of learning and consistent performance; research has also identified training an alternative behavior with positive reinforcement methods while preventing the reinforcement of, and therefore extinction, of problematic behaviors more effective then punishment. One interesting study on training methods used on dogs in obedience competitions found dogs trained with only positive reinforcement methods scored highest in obedience; dogs trained with aversive techniques scored the lowest and also had the most behavioral issues. Dogs trained using mixed methods scored between the groups. This study was particularly relevant because the dogs in the study were raised in homes, with ‘real world’ ranges of experience that makes the study’s findings more meaningful to life outside of a laboratory.
The constructional graded approach (or ‘errorless’ as described by PhD. James O’ Heare), focuses on building new desirable behaviors and alternatives to problem behaviors. This approach minimizes the use of aversive methods, and with an emphasis on positive reinforcement. It can be seen as ‘minimally aversive’, because other than positive reinforcement, everything else (negative reinforcement, negative punishment, positive punishment) causes some level of discomfort. A minimally aversive graded approach is intended to utilize the least aversive techniques necessary, breaking down tasks into small, achievable steps using added reinforcement whenever possible. ‘Prompting’, ‘shaping’, and ‘preclusions’ are three common techniques used in doing this. Basically, the trainer manipulates the environment wherever possible and breaks down the terminal behavior (the objective) into incremental steps that can be reinforced until the terminal behavior is reached. Escalating in aversiveness or reliance on other coercive or forceful tools and methods is a sign of a deficit in the trainer and/or problems with the training plan, rather than the animal.
The equestrian world has normalized the use of aversive training techniques, as can be seen in the range of training apparatuses (whips, spurs, leverage bits) and use of predominately punishment and negative reinforcement in shaping behavior. Round penning and lunging are two common ways negative reinforcement is used horse training. Unable to escape, the horse is run to exhaustion until it learns that the only way to get relief is by submission. However, the act of running a horse in circles does little to train the animal what behavior is actually wanted. It is also used as punishment when the horse does not submit. Spurs, whips, and harsh bits are equally used to compensate for lack of skill, patience, and forethought on the trainer’s part. These techniques remain entrenched in the equestrian world because although they often appear to work, they are also directly connected to stress, behavioral issues, and trauma responses, such as hypersensitivity, phobias, and learned helplessness.
              Aversive training techniques impair learning by causing an increase in the stress hormone cortisol, which reduces plasticity in the brain. They also trigger avoidance and escape behaviors, which often turn into aggression if they are unable to escape the aversive stimuli. It is difficult to avoid any unpleasantness; aversiveness can be seen on a scale from mildly annoying to severely painful. The point of the graded approach is to utilize the least aversive means possible in eliciting behavior change, generally by providing a greater incentive for the preferred behavior versus the problematic behavior. In behaviorology, extinction rather then punishment is preferred for problematic behavior, in that an acceptable alternative is rewarded, and reinforcers for the problem behavior are removed. When the problem behavior no longer functions to produce reinforcement, its prevalence decreases. Punishment used to eliminate problem behavior does not address the reinforcers maintaining the behavior and only suppresses the behavior. Without continual threat of punishment, the behavior will return full force. Not only is it ineffective, but punishment also comes with a high risk of side effects.
‘Learned helplessness’ is a trauma response observed across species, including in humans and horses. Learned helplessness occurs when an individual is repeatedly exposed to aversive stimuli that it cannot escape from; eventually, it gives up and no longer attempts evasion. ‘Depression’ is also used to describe this, as reduced effort by the subject is seen. Sadly, the symptoms of learned helplessness in horses are frequently misinterpreted as ‘well trained’, ‘desensitized’, ‘bomb proof’, and the like. It can also be seen if repeatedly enduring confusing, inconsistent and unclear training sessions.
Hypersensitivity, fear and phobias occur after a single or repeated incidents of intensely aversive stimuli. When experiencing a pain or fear reaction, the animal will often associate other stimuli present with this incident, even if they are not the direct cause of the pain or fear. As with humans, cumulative traumatic experiences increase the likelihood of post traumatic stress disorder.
              Respondent counter-conditioning using positive reinforcement to change the association if a stimulus from a negative to a positive conditioned response is an ethical way of reducing fear and undoing damage from past traumas. Rather than ‘flooding’ or ‘exposure therapy’, respondent counterconditioning with positive reinforcement gradually exposes the subject to the aversive stimuli, on a scale from least to most uncomfortable, giving the subject time to adjust without overwhelming them or eliciting an escape reaction.
In the case of fearful, depressed, or otherwise ‘disempowered’ subjects, an improved and enhanced environment, addressing any health deficits, and using a continuous positive reinforcement schedule when training can help rehabilitate the subject, rebuilding their confidence and desire to participate in training. The graded (or ‘errorless’ approach) is beneficial, as it sets the subject up for success. As with all training projects, the subject is transitioned to an intermittent reinforcement schedule as they improve in competence and confidence.
Feared dependency on treats, a commonly used unconditioned reinforcer in training, is a common reason equestrians are hesitant to emphasize positive reinforcement in training. When teaching a new behavior, the subject is originally placed on a continuous reinforcement schedule that is then transferred to an intermittent reinforcement schedule. Placing the reinforcer on a variable schedule prevents both dependency and extinction. Other concerns about training with treats include ‘disrespect’, crowding or nipping. The trainer can prevent this crowding by extending their arm to give the treat and only reinforcing if the horse does not crowd or mouth the trainer to get the treat, thereby preventing this behavior from being installed in the first place. Remember, all behaviors serve a function; if the horse wants the treat but crowding the handler causes the treat to be withheld, they will not be inclined to do so. Consistency is key… occasionally giving the treat despite crowding or undesirable behavior puts that behavior onto an intermittent reinforcement schedule, which will further entrench the behavior and make it resistant to extinction. Used skillfully, training with positive reinforcement will not create these problematic behaviors on its own. Training horses with treats is not new; it has commonly been used in dressage training and circus training.