Pup body language was the subject of the very first thing published on this site: a black poster showing ten positions, each labelled with a state of mind. Calm, anxious, relaxed, alert, playful, dominant, frightened, excited, aggressive, submissive. It has been read thousands of times, and it still works as a starting point. What has changed is the research. Since that poster went up, animal behaviour scientists have taken apart several of the readings it relies on, and three of the ten labels no longer hold.
The chart borrows its vocabulary from dog trainers, who borrowed it from ethology, which had already simplified it. Nobody sat down and legislated pup body language. It accumulated, the way a dialect does, and that is worth remembering before treating any diagram of pup body language as a standard.
Several positions on the poster are honest transcriptions of real canine behaviour. The forward lean with the head extended, the low crouch with the head turned away, the play bow with the chest down and the rear up: dogs do all of these, and they are readable. Others are human staging. A dog cannot hold a posture for a photograph, and it does not need to make its state legible to an audience. A pup does, which is the whole difference. Half of pup body language is inherited from the animal; the other half is performance, and knowing which is which saves a lot of arguing.
The poster’s “submissive” is a pup on his back, belly exposed. It is the most quoted reading in the entire vocabulary, and it does not survive contact with the data. Kerri Norman and colleagues at the University of Lethbridge analysed two sets of videotaped dog play in Behavioural Processes in 2015 and found that not one rollover could be classified as submissive. Most were combat manoeuvres: defensive, to evade a bite to the nape, or offensive, to launch an attack from below. Size difference between partners had no effect, and the supine phase was overwhelmingly brief.
So a pup on his back is not necessarily yielding. He may be setting up. Reading pup body language as a hierarchy chart is how people get that one wrong.
The classic explanation of the play bow is that it clarifies intent, announcing that whatever follows is not a real fight. Sarah Byosiere, Julia Espinosa and Barbara Smuts tested that in 2016 in the same journal, looking at what happened immediately before and after each bow in adult dogs at play. Play bows occurred most often after a brief pause. They function to restart play that has stalled, not to excuse an ambiguous move.
For a pup, that is a practical detail rather than a trivia point. It is one of the few pieces of pup body language with a specific job: the bow is what you offer when the energy has dropped and you want it back.
Alexandra Horowitz filmed 14 dogs in a setup where owners left the room and a forbidden treat was available, then varied whether the owner knew what had happened. The behaviours everyone recognises as the guilty look appeared no more often when the dog had actually disobeyed. They appeared when the owner scolded, and the effect was strongest when the dog had done nothing wrong.
Transpose that. On the poster, “anxious” and “frightened” are presented as internal states a pup arrives at on his own. Often they are responses to the handler’s tone and posture in the previous ten seconds. If a pup keeps producing them, the question is not what is wrong with the pup. A good part of what reads as pup body language is a mirror held up to the handler.
Two of the dog’s clearest signals are simply unavailable to a hooded pup, and that constraint shapes pup body language more than any poster does.
Juliane Kaminski’s team dissected dog and wolf heads for a 2019 paper in PNAS and found that the levator anguli oculi medialis, the muscle that raises the inner eyebrow, is consistently present in dogs and absent in wolves. Dogs produce the movement more often and more intensely, and the highest intensities are exclusively canine. The expression resembles human sadness and probably triggers a nurturing response, so it looks like roughly 33,000 years of selection by people who liked being looked at that way.
The irony is hard to miss. The single facial signal domestication added to the dog is the one a hood covers completely.
The finer signal is also unavailable. Marcello Siniscalchi and colleagues showed in Current Biology in 2013 that dogs watching video of another dog wagging with a left bias displayed higher cardiac activity and more anxious behaviour than when the wag was right-biased. Dogs read an asymmetry most humans never notice. A silicone tail cannot produce it.
Strip out the face and the tail and pup body language runs on three things: where your weight sits, how high you carry your chest and head, and how fast you move. That is the entire instrument. It is coarser than a dog’s, and it is also more deliberate, because every position is chosen.
This is why the poster still earns its place. Not as a translation table, but as a reminder that you are being read continuously, and that ten legible shapes are worth more than a hood you paid a lot for. Building a pup persona is largely a matter of deciding which of those shapes are yours.
A position describes where someone is right now. Pup body language reports a state; it does not authorise anything, and it never has. A pup rolling onto his back is not agreeing to what you had in mind, any more than a dog doing it is conceding a fight. Ask, then act, and keep a safeword or a clear non-verbal equivalent available for anyone who cannot speak while hooded.
Two warnings worth taking seriously. If a pup holds the low, turned-away shape for long stretches rather than moving through it, something is wrong, and continuing because the poster calls it a valid position is not an answer. And if the comedown afterwards runs into days rather than hours, that belongs in a conversation about aftercare, not in a diagram.
Want help finding your own set of positions? Contact me and we will work through it.
– Kerri Norman, Sergio Pellis, Louise Barrett, S. Peter Henzi, “Down but not out: supine postures as facilitators of play in domestic dogs”, Behavioural Processes, 110, January 2015, pp. 88-95. PMID 25217866.
– Sarah Byosiere, Julia Espinosa, Barbara Smuts, “Investigating the function of play bows in adult pet dogs (Canis lupus familiaris)”, Behavioural Processes, 125, April 2016, pp. 106-113. PMID 26923096.
– Alexandra Horowitz, “Disambiguating the ‘guilty look’: salient prompts to a familiar dog behaviour”, Behavioural Processes, 81(3), July 2009, pp. 447-452. PMID 19520245. 14 dogs.
– Juliane Kaminski, Bridget Waller, Rui Diogo, Adam Hartstone-Rose, Anne Burrows, “Evolution of facial muscle anatomy in dogs”, PNAS, 116(29), 16 July 2019, pp. 14677-14681. PMID 31209036.
– Marcello Siniscalchi, Rita Lusito, Giorgio Vallortigara, Angelo Quaranta, “Seeing left- or right-asymmetric tail wagging produces different emotional responses in dogs”, Current Biology, 23(22), 18 November 2013, pp. 2279-2282. PMID 24184108.