Verified by 
I was sitting in a hotel room and surfing the available TV channels to find something to occupy a few minutes of my time when I encountered a dog behavior show. It was hosted by a man known to believe strongly in theories of canine dominance and submission, and who frequently uses those ideas to attempt to change dog behaviors. As part of the current episode he was explaining that sometimes it is difficult to determine whether a dog is dominant or submissive if there are no other dogs living in the house. He went on to suggest that you can get a good understanding of your dog's relative dominance by watching its interactions while at an off leash dog park.
The camera cut to a scene of dogs playing and the host's voiceover continued saying something like "A dog may appear to be dominant, but sometimes in play situations you will find that same dog throwing himself on the ground with his belly upward which is a movement which is completely the opposite to any possible show of resistance or dominance." Meanwhile the image on the screen showed a large Labrador retriever roll on its back at the approach of something which looked like an Australian cattle dog who was only two thirds the height and weight of the Lab. The show's host commented that this was an example showing that the larger dog was really acting submissively and the cattle dog was really the more dominant in this pair."
As I watched this video clip I found myself disagreeing with this TV personality's interpretation of the specific behaviors on the screen. This is not simply an issue of whether the best way to interpret dog behaviors is in terms of dominance and pack hierarchies (this has been discussed elsewhere click here or click here for discussions) but rather on the simple interpretation of this rollover body signal in dogs. Konrad Lorenz, the Nobel prize-winning ethologist, was the first to describe the basic signs and signals indicating dominance or submission among canines in his 1952 book "King Solomon's Ring". He reported that when two dogs or wolves are engaged in a conflict, the defeated animal rolls over on his back and offers his neck to the other. He claimed that this serves as a signal of submission and that any dog who performs such a rollover will never be seriously bitten. This rolling over behavior has come to be universally accepted as a sign that the dog performing this action has given up the fight and is signaling that he is no longer resisting. So if we simply look at the rollover behavior presented in this TV clip that certainly is a possible interpretation of the relationship between the dogs. Still, looking a little bit more closely at the behavior, there was a significant difference from Lorenz's description of the submissive rollover. Lorenz noted, "However, this strange inhibition from biting persists only as long as the defeated dog or wolf maintains his attitude of humility." In other words for this rollover to signal submission, the dog that goes down must stay down until the other dog stops all signs of aggression. Yet that was not happening here. The Labrador retriever did indeed show a rollover behavior but it lasted only for a couple of seconds until the cattle dog approached his face at which point the Lab gave an air snap and came back to his feet and began chasing the cattle dog. Considering the whole sequence of events I certainly had no reason to believe that the larger dog's rollover behavior was indicating any kind of real submission, but rather it seemed to me that this was something having to do with play behaviors.
A play bow
There has already been research which shows that play behaviors are special, and there are certain signals that help to distinguish play behaviors from not-play behaviors. Perhaps the best known of these is the play bow, where the dog lowers its front legs flat to the ground with its hindquarters up, and this is often accompanied by a so-called "play face" with the mouth open and the tongue lolling out. This set of signals can serve as an assurance to a dog's playmate that everything being done at this time is play — so, for instance, that bite in the face that he just received was meant as play and was not part of an aggressive act.
A research team headed by psychologist Kerri Norman of the University of Lethbridge in Alberta, Canada, recently looked at rollovers during play behavior in a study published in the journal Behavioural Processes*. Their study had two parts. Part one consisted of 33 staged play sessions where a medium-sized female dog was paired with 33 new play partners of various breeds and sizes. Part two involved looking at 20 YouTube videos where two dogs played together, with half of the videos including similar sized dogs and the other half including dogs of different sizes. Their first finding was that in these dog play pairs not all interactions involved rollovers (only nine of the partners out of 33 in the staged play, and 27 of the 40 dogs in the videos).
The researchers next analyzed the meaning of the various rollovers. If they were actually submissive behaviors then, following the rollover, the dog that was down should remain down, or at least afterwards the play behaviors should decrease. Also, one would expect that more rollovers would be performed by the smaller or weaker partner. The scientists compared this to rollovers that were associated with the interactive or combative nature of play, such as when the rollover proceeded launching an attack (offensive behavior) or evading a neck bite (defensive behavior), or rolling in front of a potential partner before any interaction in order to convince the dog's partner to play (solicitation).
The researchers' conclusions were quite unambiguous. The smaller of the two play partners was not more likely to rollover than the larger dog. In addition, the investigators state, "most rollovers were defensive and only 9 of the 248 rollovers was submissive." In other words, the dogs performing the rollovers did not remain passively on their backs, but rather while in that supine position the dogs usually blocked the attempted bites of their play partner and then immediately launched bites back at them, often coming out of the roll and turning the play pattern into a chase.
So at least in this play situation, the rollover is not saying "You are coming on too strong and are worrying me" nor "I give up. You've won this round." In some settings, a rollover is certainly a signal that is associated with fear or is an attempt to defuse or prevent aggression but apparently not during play.
This new research should serve as a reminder that rolling over, like many behaviors, does not have a universal meaning. It is much like human language, where the context or setting determines the meaning. Consider the meaning of the English word "sharp". It's meaning changes depending upon the context that it finds itself in, such as: "sharp knife", "sharp dresser", "sharp mind," or "sharp tongue." So rolling over during play does not necessarily, or even frequently, mean submission. Rolling over during play is often just playful.
Stanley Coren is the author of many books including: The Wisdom of Dogs; Do Dogs Dream? Born to Bark; The Modern Dog; Why Do Dogs Have Wet Noses? The Pawprints of History; How Dogs Think; How To Speak Dog; Why We Love the Dogs We Do; What Do Dogs Know? The Intelligence of Dogs; Why Does My Dog Act That Way? Understanding Dogs for Dummies; Sleep Thieves; The Left-hander Syndrome
Copyright SC Psychological Enterprises Ltd. May not be reprinted or reposted without permission
* Data from: Norman K., Pellis, S., Barrett, L., & Henzi, S. P. (2015). Down but not out: Supine postures as facilitators of play in domestic dogs, Behavioural Processes, 110 88-95. DOI: http://dx.doi.org/10.1016/j.beproc.2014.09.001
