A Property of Relationships or an Indicator of Stress?
First published at PsychologyToday.com on May 9, 2011 at PsychologyToday.com.
One of the Good Guys
As someone who casts a critical eye on the latest developments in dognitive science, I’m happy to report that John W. S. Bradshaw has joined us here at PsychologyToday.com. In my view Bradshaw is one of the good guys, someone who’s interested in questioning the prevailing wisdom about how dogs think and learn. In fact, Dr. Bradshaw has done two studies (with N. J. Rooney) showing that playing tug-of-war—long thought to increase dominant tendencies in dogs—can actually make dogs more obedient, and less dominant! (This is my interpretation, not necessarily Bradshaw’s.) 
He’s also written a fairly compelling argument against the idea that dogs have dominant tendencies, personalities, or character traits, providing numerous examples of how the term dominance doesn’t apply to canine behavior. This paper is a must read for any serious dog trainer. (“Dominance in domestic dogs—useful construct or bad habit?” Journal of Veterinary Behavior, 2009, 4, 135-144.)
In his first outing here Bradshaw writes about our current understanding of how and why cats raise their tails when greeting other cats, people, and dogs. And in doing so questions the idea that animals are capable of understanding the concepts of rank and status.
He writes, “Biologists seem to be uncertain as to the point in the evolution of the mammalian brain where a concept of ‘status’ becomes possible—some place the divide between primates and non-primates, which would mean neither the domestic cat nor the domestic dog would qualify.”
The Concept of Dominance
However the idea of dominance itself has yet to be questioned by mainstream science in any serious way. In fact, even Bradshaw himself, in his 2009 position paper, writes that “dominance is correctly a property of relationships.” Then he proposes a theory of dominant relationships based on associative learning theory.
I don’t think either of these things is true. I’ll tell you my reasons why in a bit, but first I need to provide a little context.
In previous articles in this series I’ve discussed modern research that calls into question two of the most basic precepts of learning theory: A) that animals learn through the law of consequences  (there is compelling evidence to suggest that they learn through a feedforward system, a la Shannon’s information theory, rather than through a feedback system, a la Thorndike), and B) that the brain comes equipped with “reward circuits,” referring to neural pathways in which the feel-good chemical, dopamine, is released. There is more than enough evidence showing that these are “attentional,” not reward pathways, especially since dopamine is released when both positive and negative experiences take place. 
Yet this information hasn’t stopped most scientists from still referring to dopaminergic pathways as “reward circuits.”
This series is an attempt to unify some in the dog training world by bringing together the best aspects of the three most common forms of training, +R, pack leader techniques, and drive training, which is how police dogs, search-and-rescue detection dogs, etc., are trained. In other words, there is more than one “correct way” to train a dog.
Skinner Was Heavily Influenced by Freud
I think it’s important to remember that positive reinforcement is an experimental outgrowth of Freud’s pleasure principle which states that pleasure is not just about good things happening, but also the sudden release of repressed instinctive needs and emotions that have been “dammed up to a high degree.”
So contained within Freud’s definition we find the platform for both positive reinforcement (the addition of a pleasurable stimulus) and negative reinforcement (the removal of an unpleasant one). 
I think it’s also important to remember that the first dogs ever trained for obedience work were taught using play as the primary focal point of learning—particularly games that stimulate and satisfy their prey drive. In fact, when Max Von Stephanitz began training dogs in the early 1900s, he did so in part to imitate some of the predatory motor patterns found in wild wolves. So drive training, which operates primarily through stimulating then releasing a dog’s instincts and emotions, existed long before the Skinner box was conceived. It was also based on the natural behaviors of dogs and wolves, not on the behaviors of rats and pigeons locked in boxes in a laboratory. Plus it operates on some key Freudian principles, which have recently been validated by modern neuroscientists. 
What Is the Scientific Definition of Dominance?
There actually is no single, coherent definition. The one presented by Dr. Bradshaw in his 2009 position paper defines it as “an attribute of the pattern of repeated, agonistic [i.e., aggressive] interactions between two individuals, characterized by a consistent outcome ... [where] ... the status of the consistent winner is dominant and that of the loser subordinate’’ (C. Drews, Behavior, 1993).
Is this true? Remember, most of the ideas we have about dominance in dogs and wolves came initially from studies done on captive wolves, who didn’t know one another, and who were unable to hunt together as a group. We now know that in wild packs dominant and submissive behaviors are not as prevalent as was previously thought (Mech, L. David, “Alpha status, dominance, and division of labor in wolf packs,” 1999.)
I think there’s also a very tenuous thread leading to who’s dominant and who’s submissive in a typical interaction that takes place between the pack parents. The male brings a dead animal back to a place near the den, holding it tightly in his jaws. The female comes toward him, hoping to feed it to her pups. As she approaches the male stands tall, his body grows stiff. He grips the dead animal tighter and tighter in his jaws. The female’s posture, on the other hand, softens. She gets lower and lower to the ground. The closer she comes, the taller and more “dominant” the male’s posture becomes, and the more “submissive” the female seems to be. Then, at the last second, when the female is right up close, so low to the ground as to almost be rolling over on her back, she snatches the prize from her mate and scurries back to the den to feed it to her offspring.
The male acts dominant, yet he loses. How does that gibe with the idea “the status of the consistent winner is dominant and that of the loser subordinate” when the “submissive” female is the consistent winner here?
What About Dominance in Dogs?
In my 20+ years of working with dogs, I have only encountered one dog that I thought was dominant. And it turned out that he was being beaten routinely by his owner. That isn’t to say that other trainers wouldn’t interpret their behavior in the same way I do. But the most interesting thing about my experiences with such animals is that I generally win my “battles” with so-called dominant dogs by behaving in a non-threatening manner.
In one especially dramatic case, I was met at the door by an aggressive Rottweiler who jumped up on me, put her front paws on my shoulders, and snarled at me, seemingly ready to tear my face off. Without a moment’s hesitation I broke eye contact and softly praised the dog. She immediately jumped down and began licking my hand.
Five minutes later she was lying next to me on her owner’s couch, with all four legs in the air, happily mouthing my left hand (which I not only allowed her to do, I actually put my fingers into her mouth to give her the pleasure of releasing the oral tension she’d been feeling when I came in the door.)
She did a complete 180, behaviorally-speaking, because I broke eye contact and praised her softly when she was about to bite my face off! (WARNING: Do not try this at home!)
Do Wolves Learn via Association?
This brings up another problem. Not only does Dr. Bradshaw state that dominance is correctly a property of relationships, and provides a definition which is disproven every time a female wolf snatches a small prey animal from her partner’s mouth (or when an iconoclast like myself does things “backwards”), he also suggests that all such encounters can be explained via learning theory.
If that were true, the male wolf wouldn’t continually lose his encounters with the female. He would learn by association that her “submissive” posture is a trick. And there is no aspect of learning theory I can think of that can explain why praising a dog for acting aggressive would instantly turn a vicious, snarling dog into a sweet, adorable cupcake who happily mouths your hand instead of trying to rip your face off.
Let’s look at the behaviors of the male and female wolf again, strictly from the perspective of their body language. As the conflict progresses, the male’s body becomes stiffer and stiffer. The female, meanwhile, becomes softer and more flexible. Stiff postures are caused by tension. Soft, flexible postures are marked by an absence or release of tension. It stands to reason then, that what scientists call dominant tendencies, personalities, and character traits are products of tension or stress, and can be resolved (in some cases quite quickly) simply by reducing a dog’s tension and stress in ways that may seem impossible, or at the very least, counter-intuitive).
This also shows that dominance is not necessarily “a property of relationships,” it may be an indication that an animal is simply experiencing internal tension or stress.
If Dominance Is Caused by Stress, Let’s Test That
My initial experience with the female Rottweiler is one I wouldn’t care to repeat. Yet such a one-time incident doesn’t provide empirical proof that praising a dominant dog can reduce her aggression. In order to prove such a thesis, we’d have to show a pattern of repeatability; we’d have to set up a study, where this thesis could be tested with dozens of dogs, over a substantial period of time.
This could be done fairly easily, and safely, as long as we avoid situations like the one I put myself into, where the successful outcome may have been based largely on dumb luck.
If dominant behaviors are the result of stress, that’s easily provable, empirically, because we can measure stress through serum cortisol levels. In fact, stress levels can now be measured over time, rather than by testing urine or saliva, using a new technique that measures cortisol levels in hair.
Another possibility—though a futuristic one—has been suggested by a new device invented at Brookhaven Labs/Stonybrook University to do PET scans on rats, in situ. It’s a ‘wearable’ device which allows for simultaneous study of behavior and brain function in animals. It’s currently only used on rats, and it requires an injection into the rat's bloodstream of radioactive tracers. But in the future we could have something similar to portable PET scans, portable fMRIs and all sorts of Tom Swiftian devices at our disposal. (Update: researchers at Stanford University have developed a new, non-invasive brain imaging technique that allows scientists to visualize cerebral vasculature and blood flow in mice in remarkable detail, 8/8/14.)
Meanwhile, it’s a simple formula: when you reduce a dog’s stress you cure her behavioral problems. And—as Rooney and Bradshaw have shown—it’s eminently testable.
“Life Is an Adventure—Where Will Your Dog Take You?”
Join Me on Facebook!
Follow Me on Twitter!
Link With Me on Linkedin!
1) For my own purposes I have generously interpreted these studies to mean that playing tug reduces a dog’s “dominant” and/or aggressive tendencies. I haven’t undertaken this interpretation lightly, either. It’s based on years and years of experiences with dogs who had been diagnosed as having dominant personality traits, temperaments, or tendencies, but discovered that when an owner or trainer engaged such dogs in games of fetch and tug, all such traits and tendencies disappeared.
2) In his article, “Deconstructing the Law of Effect,” Neuroscientist Randy Gallistel of Rutgers poses the problem of learning from a modern perspective, contrasting Edward Thorndike’s model, which operates as a feedback system, and a feedforward model based on Claude Shannon’s information theory.
3) In his paper “Dopamine and Reward: Comment on Hernandez et al. (2006),” Gallistel writes, “In the monkey, dopamine neurons do not fire in response to an expected reward, only in response to an unexpected or uncertain one, and, most distressingly of all, to the omission of an expected one.” [Italics mine.]
4) Freud gets a bad rap, but he was a neurologist before he became a psychologist. Freud also predicted the discovery of endorphins, dopamine, and other neurochemicals, and said that if his theories were to hold up they would have to be validated by advances in our understanding of how the brain works.
Here are two studies showing that a key Freudian concept, relating to impulse control, operates in both humans and dogs.
“The default-mode, ego-functions and free-energy: a neurobiological account of Freudian ideas.” R.L. Carhart-Harris, K.J. Friston, Oxford Journals, Medicine, Brain, Volume133, Issue 4, Pp. 1265-1283.
“Common self-control processes in humans and dogs,” Holly C. Miller et al, Department of Psychology, University of Kentucky, Lexington.