Negative reinforcement is a commonly used technique in horse training. Known more informally as ‘pressure and release’ training, those using it apply it with varying degrees of subtlety and finesse: some trainers use it in ways that appear to cause horses visible distress, while others do so in ways that appear to not cause horses distress at all.
Despite the latter use of negative reinforcement during horse training, some people feel that negative reinforcement is always unpleasant for horses and thus its use should be avoided as much as possible. Why do they believe this? While it could be their own personal biases, it could also result from what they have learned from a textbook.
Some textbooks teach that negative reinforcement is a process whereby an animal learns to avoid an aversive event by responding to a warning stimulus that occurs before the aversive event. After the warning stimulus has been paired with the aversive event enough times the warning stimulus itself becomes a conditioned aversive - something that the animal learns is unpleasant. By this logic, even a light rein cue (the warning stimulus) would become a conditioned aversive if paired with pressure on the rein (the aversive event) to teach a horse how to turn - no matter if that pressure was light or very hard. However according to research discussed in Dr Paul Neuman's talk, 'What the textbooks don’t tell you about negative reinforcement', at the recent Art and Science of Animal Training (2020), this isn't the case at all.
Dr Neuman explained that research not being discussed in such textbooks presents a very different picture: the warning stimulus does not become a conditioned aversive stimulus; it becomes a discriminative stimulus - something that simply tells the animal that their behaviour which potentially follows will be reinforced. In the case of our horse example above, the light rein cue becomes a discriminative stimulus, telling the horse that their behaviour of turning will result in no additional pressure being added. To further explain this, while discussing research involving rats who received a warning (a tone or light) before receiving a shock, Dr Neuman pointed out that such advance warning stimuli made the rats wait longer to avoid the shock. If these warning stimuli were truly conditioned aversive stimuli the opposite would have occurred: the rats would have responded as soon as the lights/sounds began.
When we are training animals it is critical to ensure that we are minimizing causing the animal to feel pain or unpleasant emotions such as fear during training. Negative reinforcement can indeed be used in ways that can frighten horses, or cause them to feel pain. Examples of this can be readily seen as you scroll through horse training videos posted by some professional trainers on Facebook. However, negative reinforcement can also indeed be used in ways that don't result in the horse feeling fear or pain. Just as with the 'other' reinforcement procedure, positive reinforcement, negative reinforcement can be a valuable tool in our training toolbox, when we know how to use it appropriately.
Dr Neuman's talk has certainly piqued my interest in diving a bit deeper into this research over the next while. Equine Clicker Training - Katie Bartlett has done a great job of summarizing what was a fascinating - but very technically heavy - presentation. Thank you, Katie. You can find Katie's summary of this talk here:
(This post was made on my Facebook business page on February 29, 2020. I think it is worth resharing here, as it remains relevant today.)