Getting Back to Basics: Pavlov and Freud.
Let’s Be Realistic About the Problems With Learning Theory
Originally published in slightly different form on January 28, 2011 at PsychologyToday.com.
Back to Basics
Patricia McConnell—a figurehead of the “positive training” movement—recently wrote a rebuttal to one of my PsychologyToday.com articles (about the dangers to teaching formal obedience skills to puppies before their brains, bodies, and emotions are ready).
My Unified Dog Theory series is an attempt to help end the divisiveness in the training world, and also to help educate dog owners and trainers that there aren’t just two training methods: the pack leader model and positive reinforcement. There’s a third, called drive training, which has been used to train working dogs—police dogs, herding dogs, search-and-rescue dogs—for far longer than either of the other two models.
Some might argue that drive training is merely an admixture of classical and operant conditioning. But I think a pretty strong case can be made for the idea that it operates more along Freudian, rather than Skinnerian or Pavlovian, lines.
Remember, the concept of positive reinforcement is really a clinical outgrowth of Freud’s pleasure principle, the difference being that for Freud pleasure was not just the addition of an enjoyable stimulus (or “reward”), but also the sudden release of internal tension or stress. And that’s essentially how drive training works.
Ontogeny Recapitulates Phylogeny
According to authors Marc Bornstein and Michael Lamb (Developmental psychology: an advanced textbook, 1999, p.20.), Freud’s interest in Darwin began when he studied biology in his first year of medical school. Freud found the biogenic theory—ontogeny recapitulates phylogeny—particularly fascinating. And although that theory eventually proved not entirely tenable, a modern version still holds as a rule-of-thumb: if a certain morphological structure pre-dates another in evolutionary terms, it also tends to appear earlier during an oranism’s embryonic development.
The forces of Nature and evolution tend to iterate, and reiterate forms and patterns, not just in the development of the morphological traits of an organism but in their emotional and cognitive abilities as well. Example: the thumb of our pre-chimp, pre-human ancestors slowly morphed into the fully-opposable thumb of modern man.
This process is true of many cognitive abilities as well; they co-opt or base themselves on already functioning abilities that lie lower on the psychological scale. For instance, the process of pattern recognition is an evolutionary pre-cursor to logic, language and math. You can’t talk, reason, or learn multiplication tables without it.
In Freudian psychology, “Developmental change is viewed as qualitative and stage-like, proceeding through tension resolution from one stage to the next.” (Bornstein & Lamb)
Each of these developmental stages are important. With puppies—just as with growing embryos—you can’t skip ahead or move things along too quickly or you’ll damage the organism in some way. Nature and evolution have put those stages there, in sequence, to serve a purpose.
There is plenty of evidence to support the idea that pushing a human child to go through a natural developmental phase too quickly may result in learning deficits, emotional trauma, and psychological damage. Freud may have been off-base about some things, but not this.
“What does that have to do with puppy classes,” some might ask. “And what harm can they do if you’re using positive reinforcements?”
In “Beyond the Pleasure Principle (1920)”, Freud wrote: “We have no hesitation in assuming that the course taken by mental events is automatically regulated by the pleasure principle. We believe, that is to say, that the course of those events is invariably set in motion by an unpleasurable tension, and that it takes a direction such that its final outcome coincides with a lowering of that tension—that is, with an avoidance of unpleasure or a production of pleasure. In taking that course into account in our consideration of the mental processes which are the subject of our study, we are introducing an ‘economic’ point of view into our work; and if, in describing these processes, we try to estimate this ‘economic’ factor in addition to the ‘topographical’ and ‘dynamic’ ones, we shall, I think, be giving the most complete description of them of which we can at present conceive.” (The Freud Reader, 594-595).
The reduction of internal tension in a dog’s mind and body is an actual, physiological process which can be physically measured in any number of ways, in real time, while the process of positive reinforcement cannot; it’s more of a statistical probability, measured after the fact. This means that physical objects, rewards, events, etc. can only be classified as reinforcements if they increase an organism’s tendency to repeat a behavior. And the fact that many puppies, who are put into obedience classes too early, forget everything they’ve “learned” once they reach adolescence indicates that their behaviors in puppy class were not really positively reinforced after all, because the learning faded over time.
Training as a Lifelong Process?
This brings up the belief that dog training is a lifelong process. In their responses to McConnell’s article, many of her followers insisted that training continues from the moment you bring your puppy home up until the dog’s last breath. This isn’t true. With most of the dogs I’ve trained I’ve found if you wait until the dog is emotionally ready, and you use the dog’s prey drive as the primary focal point of learning, then at some point, no further training is necessary. None! The dog knows his stuff and never falters.
Then there’s the problem of whether or not operant conditioning is a complete and/or entirely valid model of learning. Dr. McConnell is convinced that it is, despite the fact that over the past year or so she’s had to give up two of her dogs—both purchased as puppies from reputable breeders—essentially because she couldn’t figure out how to train or condition them get along with her older dog, Will, a border collie, whom she also purchased and raised as a pup.
”Dog Training Is Not Working That Well”
Ian Dunbar, another figurehead of the +R movement seems to have come to the conclusion recently that “dog training is not working that well,” and that certain aspects of operant conditioning are “hopelessly complicated.”
The problem may be that operant conditioning, despite its effectiveness in laboratory conditions, is just a model rather than an actual process. Pavlov didn’t think it was a real process. Some modern researcher—like Randy Gallistel of Rutgers—are now saying that while classical conditioning is real, operant conditioning may not be. Their research shows that learning may actually take place, not through the law of consequences (a la Skinner and Watson) but through a simple process called pattern recognition.
In a previous article here I gave evidence that dopamine circuits in the brain, long thought of by behavioral and cognitive scientists as “reward pathways” are actually attentional pathways, switched on by both pleasant and unpleasant experiences, as if telling the brain, “Pay attention to these patterns. Remember to do this, remember not to do that!”
Of course there are times when the process of operant conditioning—via external rewards—coincides with the reduction of an animal’s internal tension. So I’m not asking Dr. Dunbar or Dr. McConnell, or their followers, to abandon the use of operant conditioning techniques altogether, just to at least begin to entertain the possibility that they don’t hold all the answers.
Dogs have been working animals for thousands of years. It’s only since the advent of behavioral science-oriented puppy classes that dog training—in Ian Dunbar’s words—“isn’t working that well.”
So the question becomes, should our pet dogs be trained as if they’re no different than rats learning to run through a maze, or as if they were working dogs whose instincts and emotions need to be put to use?