Are We Habituating to Inaccurate Terminology within the Field of Equestrian Science, and is this Overshadowing Our Progress?

Horse trainingOne of the great advantages of a scientific approach to behaviour is that we can define specific words, ensuring everyone knows what everyone else is talking about.  Unfortunately, in the field of equestrianism, well established terms relating to learning are often used inaccurately. This may be due to ignorance, or possibly in an effort to put a positive spin on favoured training techniques relying on aversives. Even among leading practitioners within the equestrian science field, terminology is being used that is sometimes at odds with definitions that have been in place for 50 years or more. 

Implications of the Inaccurate use of Terminology:
So why does any of this matter?  Is concern about inaccurate terminology a matter of semantics, or does confusion regarding terminology inhibit our understanding of equine learning, reduce the effectiveness of our training, and compromise the welfare of our horses?
Traditional horse training relies heavily on aversives to modify behaviour. This may be necessary to some extent considering our expectations of horses. However, if we are to rely on aversives in horse training, it is critical that we are clear about what learning process are occurring when we employ specific training techniques, so the benefits and risks can be discussed objectively.  Consistency in the use of terminology is essential in order for this objective discussion to occur, and confusion has the potential to interfere with our effective and humane training of horses.  Three terms are discussed below that highlight this issue.
“Positive reinforcement”:
Positive reinforcement is an operant conditioning process where the linking of a behaviour with a stimulus increases the frequency of that behaviour in the future.  For example, if, as a horse is walking into a horse trailer, it receives a carrot, and then the behaviour of entering the trailer becomes more frequent in the future, this behaviour is said to have been positively reinforced
However, there is a recent trend to describe the removal of pressure as a “reward” or “positive reinforcement” because the horse desires it. The fact that something aversive needs to be applied for its removal to be desirable is an ugly truth that is often swept under the carpet by this inaccurate use of the term.  This type of reinforcement is actually negative reinforcement which, as with positive punishment, relies on the judicious application of something aversive to the horse in order to modify behaviour.
Habituation is a non-associative learning process where a reflexive response to a stimulus gradually diminishes as a result of tolerable exposure. Often the term is used when discussing the reduction of a fear response.  For example, horses living near a railway line, and who are able to escape sufficiently far away that they are not overly fearful, will flee progressively shorter distances and become less fearful of passing trains; they become habituated to the trains.
The ability to manage exposure to fear-eliciting stimuli so that fear is kept at a tolerable level is critical in order for habituation to occur.  Without the means to move away, a horse can be pushed to a level of fear that interferes with the habituation process.  This often occurs during poorly executed attempts at flooding1.  Flooding is initially highly stressful for the horse, and it is possible for the horse to become more fearful (sensitised) to the stimulus rather than habituating to it.  In some circles the term “habituation” is used to describe such training, which masks the risks of taking a horse past the threshold at which it can cope. 
Overshadowing refers to the classical conditioning phenomenon that occurs when two neutral stimuli are presented simultaneously before exposure to an unconditioned stimulus2. After repeated pairings the more salient3 of the two neutral stimuli is more likely to become a classically conditioned stimulus, eliciting a conditioned reflexive response. The more salient stimulus is said to have “overshadowed” the less salient stimulus during the acquisition phase of the classical conditioning process. Overshadowing can also be used to describe an operant conditioning phenomenon when the more salient of two discriminative stimuli4 exerts greater stimulus control on the target behaviour5.
More recently, the term “overshadowing” has been used to describe a process where a horse is forced through aversive means (or via operantly conditioned behaviour trained previously through the use of aversives), to remain in the vicinity of something it finds threatening.  For example, pressure on the reins, or the threat of a whip is said to “overshadow” a fear-inducing stimulus such as clipping, or a veterinary procedure. Much as someone with arachnophobia would eventually get used to spiders if they were forced into an area infested with tarantulas and fenced off with barbed wired and electric fencing, a horse may get used to the clippers if it is forced to stay near them long enough.  However - using the spider analogy - the arachnophobic person might become so panicked that they cut and shock themselves trying to escape, and the experience may leave them more fearful of spiders.  The same can happen with a horse if it is forced into close proximity to something it finds highly threatening. 
While the process of forced exposure can result in a long term reduction in fear if performed for long enough that a horse starts to relax6, this is a flooding through response prevention process, rather than overshadowing.  The use of the term “overshadowing” again masks the risks of flooding: sensitisation, the development of dangerous escape strategies, learned helplessness, conditioned fear of the trainer or training environment, and the subjectively aversive nature of the process for the horse. 
While it is exciting that the principles of learning theory are finally trickling through into professional horse training circles, there are still discrepancies between the ways certain learning theory terms are used and their well-established definitions. Without clarity within academic circles regarding terms relating to associative and non-associative learning processes, and the practical application of these processes, we risk retarding an opportunity to bring about meaningful training and welfare improvements in the wider equestrian world.

1  A process where an animal is exposed to a highly fear-inducing stimulus without any ability to escape.
2 A stimulus that causes an instinctual involuntary response by the animal.
3 Noticeable
4 Cues
5 For example a “stand” hand signal and “stand” verbal cue may be used together during training a horse to stand still, but after training the horse is more responsive to the hand signal than the verbal cue (leading to a lack of response to the verbal cue that the trainer thinks is in place).
6 The length of time needed before the horse relaxes will depend on the subjective prception of threat experienced by the horse.