Publications on dog training initially began distribution in the 1800’s and were centered around many of the same fundamentals we use today. There have always been a variety of commonly used approaches, however their individual effectiveness has only been put to more challenging analyses in the past 20 years due to the explosion of dog related research in academia and the rise of the internet as a tool for communication of information. While many hybrid iterations fall on the spectrum between or near Positive Reinforcement and Dominance Theory, these concepts are the basis for almost all modules of training. Below we outline our thoughts on what the Modern Approach should be, based on the research and knowledge accumulated in the field over time.
Positive Reinforcement
The idea of positive reinforcement in dog training predates modern dogs themselves. One theory of how dogs were originally domesticated states that it is possible that wolves who were docile enough to be near or even approach people could get scraps of food the people did not eat, while aggressive individuals were likely killed. Though extreme, this removed aggressive individuals from the population and allowed for more “agreeable” wolves to breed going forward, providing an evolutionary advantage to wolves who could better co-exist with humans. In essence, wolves who could behave copacetically with humans would be rewarded with food, further cementing positive behaviors in those individuals and, eventually, their offspring.
Positive training techniques appeared in print as early as 1848 when the book Dog Breaking instructed, “Caresses and substantial rewards are far greater incentives to exertion than any fears of punishment.” Clicker training, which emphasizes learning by association, was first described by B.F. Skinner in the paper ‘How to teach animals’ published in the Scientific American in 1951 and expanded upon by Keller and Marian Breland shortly thereafter. However, the theory first found success as a widespread training method for marine mammals in aquariums. This did not become widely used in dog training until Marine Biologist Karen Pryor connected it to the learning theory of classical conditioning famously coined by psychologist Ivan Pavlov. Pavlov noted a dog’s natural response to food was to begin salivating. Over time he noticed that the dogs began to salivate before food was present but only when they heard the food cart coming. To investigate this phenomenon, Pavlov performed his famous experiment in which he introduced a bell, a stimulus that results in no response (known as a neutral stimulus) before feeding the dogs. After some time passed, the dogs would begin to salivate at the sound of the bell whether or not food was present. With this in mind, Pryor noted how clicker training differentiated from traditional training methods by allowing learning-by-association to occur and natural reactive behaviors from an animal to be refined (aka “shaping”) from there. Other advancements to the field from these discoveries were that learning should occur in brief sessions and contain variety to maintain the animal’s interest, and that clicker training did not require any type of punishment to effectively teach desirable behaviors.
Dominance Theory
The concept of dogs and humans vying for social hierarchy dates back to Konrad Most’s study of military oriented working dogs and the misconception that behavior theory of the modern-day dog should relate to that of wolves seen in the wild. Dominance theory gained popularity in the mid-1900’s and peaked in the early 2000’s when it was highlighted on Caesar Milan’s popular TV show “The Dog Whisperer.” Recently, this method has received significant pushback as its techniques have been largely rebuked as an optimal method by academic studies and modern-day experts. In many cases this has been found to bring out harmful behavioral side effects related to an anxious, unsafe state of living in the world for a dog. We are not fans of Dominance Theory in general, however we believe the majority of problems resulting from it are because the method has been misunderstood and over applied. Some aspects related to dominance theory may have a place in dog training, but these situations are rare and do not apply to everyday reality with companion or family dogs. We recommend if people choose to incorporate aspects from this theory of dog-training that it be done infrequently and with the help of an experienced positive reinforcement-oriented trainer. In human parenting studies, at least a 6:1 ratio of positive praise to negative reinforcement is required to prevent harmful side effects of negatively shaping behavior. We believe that’s a meaningful ratio for dogs too.
Modern Theory
It’s become clear that canines are complex animals. Neither Positive Reinforcement nor Dominance Theory adequately explain the wide range of behaviors and skills dogs exhibit. Our dogs have empathy; can think independently; are capable of assessing what others think; can consider potential rewards and new experiences; and have significant memory capabilities, among other complexities. Domestication in tandem with humans means that having such a special relationship over time has also changed them. Thus, we think it’s appropriate to model similar parenting theories for humans with adjustments related to dogs’ abilities. In general this includes a heavy emphasis on positive reinforcement as a basis for all activities. Making training fun or employing a “play to train” mentality is an important building block.
Where this diverges from the classical thoughts about positive reinforcement is in the acknowledgement of canines as independent thinkers who make choices based on a variety of factors. Dogs will manipulate, try to problem solve through trial and error, and learn from observation. Therefore, we place more emphasis on choice-based training to encourage their confidence to make decisions independently, have curiosity, and place them in situations where independent problem solving skills are developed starting at an early age. This type of learning correlates an increased neural aptitude with having an effective line of communication with your dog that is reaching heights most people never could have imagined. Additionally we need to set boundaries, and with some particularly determined breeds, solely practicing positive reinforcement may not work best. We recommend following the ideal praise-to-correction rate studied in humans at a bare minimum of, and hopefully higher than, 6:1.
Corrections should require as little physical or verbal energy as possible from us, and are often most effective when mimicking the energy level you desire. As shown below, the positively-oriented, minimally corrective style was most effective among methods in the study, while purely positive methods were second most effective and negatively oriented methods (like an e-collar) were seen to be least effective. Another example to support these results shows up in Service Dog training wherein purely positive methods can often result in dogs not knowing what to do without constant guidance from a human, resulting in anxiety for both parties. Incorporating choice-based exercises to develop problem solving skills into learning instead of solely association-based is effective approach to avoid this problem. It does so with the intent of providing a healthy environment and stimulation for the brain to develop, while respecting and bringing out the advanced intellectual and emotional capabilities inherent to our dogs.
