User:Thesoxlost/Dog Training

This page is a work in progress for an article on dog behavior and training. At this point, it is mostly notes, and it may never pass WP:OR criteria.

Schools of dog training have been strongly motivated by psychological models of human and animal behavior and learning. Although there are myriad training paradigms, there are far fewer academic frameworks from which these paradigms are derived.


 * Evolutionary Psychology:
 * Behaviorism:
 * Learning Theory: Learning theory grew out of Behaviorist psychology, and it likely means different things to different academics and trainers. It is advanced beyond Behaviorism because it acknowledges the existence of internal states and their importance in guiding learning and behavior. For instance, a central concept to learning theory is reinforcement learning. At the theoretical level, reinforcement learning is a formal framework for learning from feedback.  According to this framework, learning can be achieved by actively anticipating the reward that one will acquire.... Dopamine... Wolfram Schultz.

Brief history of psychology
Much of 20th century experimental psychology stemmed from or is strongly linked to animal learning. This relationship began with Charles Darwin, by putting human psychology on the same level as animal psychology, and by proposing a framework by which psychological factors--amongst all other genetic factors--evolve. Prior to Darwin, the study of animal behavior was a philosophical pursuit distinct from the study of the rational human mind. By theorizing that human psychology evolved according to the same framework by which other species evolved suggested that scientists could better understand the complicated system of human psychology by investigating simpler animals. Early comparative psychologists attempted to characterize the traits that certain species had developed (very often in dogs), although in a non-scientific manner. George Romanes and C. Lloyd Morgan are examples of early psychologists who sought to understand canine psychology by cataloging their behaviors.

Morgan set the foundation for the behaviorist movement that was to come by arguing for a "Law of Parsimony," which states that behaviors should not be explained in terms of complex mental processes if a simpler explanation would suffice. Thus, from this standpoint, animals cannot be attributed thoughts or emotions unless simpler models fail to explain their behavior. Morgan's canon still influences contemporary psychological models, although nearly all psychologists believe that models that do not include complex internal representations do fail to explain animal behavior.

The next paradigm shift in psychology came in the form of the behaviorist revolution. One of the fathers of behaviorism, John B. Watson, argued that psychology had become too subjective, and that psychology needed an objective medium that was accessible to and observable by all researchers: behavior. According to this school of psychology, internal representations, such as thoughts, feelings or emotions, are not only overly complex (violating Morgan's canon), they are also unobservable and unscientific; their existence is impossible to test: although we can observe a dog panting, we cannot observe the construct of "thirst," and thus there is no scientific basis to say that a panting dog may be thirsty.

A great deal of our understanding of psychology--and many of the methods by which we train dogs--stem from Behaviorism. The earliest form of reinforcement learning was framed by Behaviorists, notably Ivan Pavlov and B.F. Skinner, in the form of conditioning. Conditioning is not a theory, nor a form of learning; rather, it is a procedure for affecting animal behavior by associating a reward with a stimulus. In classical conditioning, animals passively experience a stimulus (e.g., the ringing of a bell) followed by a reward (e.g., food). Over time, the animal learns to anticipate the reward when it experiences the stimulus. In instrumental conditioning, animals learn that an action (e.g., pressing a lever) results in a reward. Over time, this learned association reinforces the behavior. Thus, in the original framework, a hungry dog who has learned to hit a lever for a treat doesn't hit the lever because it is hungry and wants food. "Being hungry" and "wanting food" are internal, unobservable constructs. This behavior can simply be explained as reinforcement: the repeated rewards associated with the lever press bias the animal to make that behavior reflexively. Much of contemporary dog training applies this understanding of canine psychology by rewarding a dog for good behavior and punishing it for bad behavior. Although these methods work, behaviorists were observing only the tip of the iceberg--much of animal behavior cannot be explained by reinforcing a given behavior in response to a specific stimulus. This discovery led to the cognitive revolution.

Cognitive psychology emerged in response (or, rather, as a reaction to) behaviorism. This school emphasized the importance of internal, unobservable variables in understanding behavior. One of the fathers of cognitive psychology, Edward Tollman, noted that behaviorism required a huge number of connections between stimuli and responses, and argued that models that include intervening variables are simpler than those that do not. For instance, by assuming that an animal has internal variables, such as fear, which are not directly observable, nonetheless reduces the complexity of the relationship between fear-inducing stimuli and fearful responses. As a result, cognitive psychology focused not on the simple association between stimulus and response, but rather focused on the internal computations (termed cognition) that best explained animal behavior.

This brief history of psychology is important for putting models of animal learning and behavior in perspective. The applied field of dog training is largely rooted in the early studies of animal learning. The academic field has progressed a long ways from these original theories, but they have largely failed to make their way into dog training. Claims of academic validity by association with behaviorism implies a scientific understanding that has long been abandoned. I would argue that the most complete picture of dog psychology comes from a contemporary understanding of learning theory combined with an understanding of canine evolution and ethology.