Reading Dog Body Language

Breed decides where a dog’s neutral ear position is. Some dogs show the whites of their eyes at rest, and whale eye is a term from practice that has not been measured in a study. Lip and nose licks show up in calm and tense situations alike. Raised hackles show arousal without saying what kind. Freezing is common between dogs and is often mistaken for calm.

Airplane ears, the different styles of wag and trembling come from clinical descriptions and trainers’ experience more than from studies, and researchers rarely measure how a dog wags. In play, rolling onto the back is usually a move in the game, and research on play has not found it to be a sign of surrender. Whining also rose during happy greetings in one study. What we know about barks comes mostly from how people rate recordings: listeners hear low, harsh, fast barks as threatening, and dogs bark lower when a stranger approaches.

In a survey of nearly 1,200 owners, trembling and whining were the stress signs people named most. Looking away, turning the head, yawning and nose licking were rarely mentioned. In videos of dogs reacting to ordinary household noises, the most common reaction from the people filming was amusement. When adults watched videos of anxious dogs with small children, most of them rated those dogs as relaxed.

Context can also mislead. When researchers edited footage so the same dog appeared in different scenes, people’s ratings of the dog followed what the person in the scene was doing, rather than what the dog had actually experienced. Use the setting to help explain what the dog is showing, and base the read on the dog.

Ethologists named these displacement activities, meaning everyday behaviors that appear in tense situations and are thought to reflect conflict or arousal. In dogs, lip licking with no food present has been linked to negative or uncertain situations in several studies, and shaking off and yawning appeared under mild stress in laboratory work.

No single one proves a dog is uneasy. Dogs yawned more with their owner present in one study, shook off more when greeting an owner after a long absence in another, and sniffed the ground as often during a calm approach as a threatening one.

Sudden scratching comes from what trainers see in practice, and I see it often: a stressed dog stops to scratch when nothing is itchy. Research has barely looked at it in dogs. The one study that counted it filmed dogs as a stranger walked up to them, which is a single short moment and not the kind of ongoing stress where trainers usually see it. Trainers also list stretching and sneezing, which have not been studied.

In the same study, nose licking and sniffing the ground happened just as often whether the stranger was calm or threatening.

A study of Labradors waiting for a reward that sometimes never came looked at facial movements one at a time. Ears pulled back appeared in most of the frustrated dogs and in many of the happy ones. Nose licks were uncommon, but when they appeared they went with frustration. The authors concluded that no single facial movement identifies a dog’s emotional state reliably.

Hormone studies point the same way. In a laboratory study, the behaviors dogs showed under stress did not match their cortisol or heart rate. In a shelter study, only one behavior out of everything recorded matched the stress hormone. A review of research on positive emotions in dogs, written by two employees of a pet-food company, concluded that no single behavior reliably shows a dog is happy.

Dog faces also work differently from human faces. When researchers compared dogs and people in the same emotional situations, the facial movements were different in every one, so reading a dog’s expression as if it were a person’s leads to mistakes.

A common explanation says dogs bow to show that a rough move was meant in fun. That idea came from a study whose sample was mostly puppies a few weeks old. In a study of adult pet dogs, bows usually came after a pause in the game and were followed by chasing, running off or moving together, so in adults a bow tends to restart play. Dogs also send play signals almost only when the other dog is looking at them, and when the other dog is not looking, they nudge, paw or bark to get that dog’s attention first.

You may also hear that healthy play is evenly matched. The research on this followed puppy littermates, and their play was uneven from the start and became more uneven as they grew. What I look for is whether both dogs keep choosing to come back. A dog who keeps leaving, hiding or getting pinned has stopped playing, whatever the other dog is doing.

In a study of 60 research dogs, trained observers’ overall impression of each dog separated pleasant situations from unpleasant ones better than any single body sign or stress hormone did. A pet-food company funded the study and employed every author, and the observers knew which situation each dog was in.

The head turns, lip licks, yawns and slow movements in this section are often called calming signals. The term comes from Turid Rugaas, a Norwegian trainer, who described them as messages dogs use to calm each other and people. Her list is useful, but the evidence supports a narrower account of what the signals do.

The one study designed to test the calming idea predicted that, if these behaviors work to calm someone down, they would increase when a dog was threatened. They did not increase. Instead, dogs who did not bark or growl at the approaching stranger showed more blinking and nose licking, whether the stranger acted threatening or neutral.

A pilot study of dogs meeting each other found that when one dog responded to another dog’s aggression with these signals, the aggression eased more often than it escalated. Dogs who simply moved away got the same result, and the study had no comparison group.

The research supports reading these signals as a sign of a dog who is uncomfortable and has not escalated, which is why they are worth noticing early. Appeasing, in the table above, is described the same way. Trainers use the word for the low, wiggly, face-licking approach a dog makes toward a person or another dog, and it describes what the body is doing. Whether the dog does it to calm anyone has not been shown.

In one German sample, about four in five dogs showed fear during a standard veterinary exam. In laboratory work, a very low, crouched posture went with the strongest stress response, while lip licking, shaking off and, to a lesser extent, yawning appeared under milder stress. In another study, dogs who froze while being handled showed a heart rhythm linked to stress.

Having the owner there helps somewhat. When owners petted and talked to their dogs during a vet exam, the dogs’ heart rates and attempts to jump off the table were lower than when owners stayed silent, although the exam raised both in each group. In another trial, dogs whose owner stayed in the room vocalized less and yawned more.

Trainers and veterinary behaviorists use the distinction between fear-based and confident aggression every day, and I use it with clients. It comes from clinical description rather than from a study comparing the two. A board-certified veterinary behaviorist describes the defensive dog as having ears back, tail tucked, head down, gaze averted, weight shifted back, and lips pulled back with teeth showing, sometimes with hackles up.

Dogs who learn that aggression works can move from defensive postures to confident ones, and return to defensive postures when overwhelmed. Another veterinary behaviorist notes that a dog in a confident posture can still be acting out of fear. Confidence is also different from "dominance," which describes a relationship between two particular dogs rather than a trait of one dog.

The ladder of aggression is a teaching model from veterinary behavioral medicine that puts warning signals in order from mild to serious. Here is a simplified version.

• 01
  Small signals

  Yawning, lip licking, head turning, blinking.

• 02
  Turning or moving away

  Body turned aside, walking off, crouching.

• 03
  Stiffening

  Freezing, a hard stare, a closed mouth.

• 04
  Growl, snap, bite

  The signals most people notice.

Dogs do not always follow this order. They skip steps and often show several signals at once. One study has compared the ladder with real bites by coding 56 online videos that showed an interaction from the start through the bite. Stiffening, a low or awkward body, staring and head turning became more common across the videos from about 20 to 30 seconds before the bite, which fits the ladder. Lip licking, paw lifting and sniffing did not follow a clear order.

The 20 to 30 seconds describes when the signals became more common across all the videos, so an individual dog may give much less warning. In the same videos, people standing over the dog became more common from about 35 seconds before the bite, and people petting or holding the dog from about 21 seconds before it.

People listening to a single growl could not tell whether the dog was playing or serious, but they could tell from the rhythm of several growls in a row.

In one clinic sample of dogs referred for biting a child, two-thirds of the dogs with a known history had never bitten a child before, and most had anxiety or other behavior problems beforehand. In a trauma-center study, young children were most often bitten in the face by a dog they knew, indoors, during a friendly interaction the child started with a dog who was not moving at the time.

Researchers now measure frustration as its own trait in dogs. Owners describe it partly through signs usually linked to stress, such as lip licking, yawning, mounting and shaking off, so the same lick can come from worry in one dog and frustration in another.

In interviews with trainers who work with dogs who react on leash, fear and frustration were the emotions they named behind it. The two can look much the same at the end of a leash, and in my experience the only thing that truly tells them apart is context.

Researchers describe tail wagging as one of the most visible and least measured dog behaviors, and it shows up in high-arousal states as well as happy ones.

You may have read that wags lean right for something the dog likes and left for something unsettling. That finding came from a lab study that tracked tail angle on video, and a follow-up found that dogs watching left-leaning wags on a screen had higher heart rates. The study’s senior author has said the difference is hard to see outside the lab, so it is not something to look for on a walk.

In one experiment, owners left the room and were then told, sometimes falsely, whether their dog had eaten a forbidden treat. The guilty look depended on whether the owner scolded the dog, and it was strongest in dogs scolded when they had not eaten the treat.

In a second study, owners were asked not to scold, and those guessing from the greeting alone did no better than chance. A pet-food company donated a camera and treats for that study. Taken together, the two studies show the look follows the owner’s reaction, so it cannot tell you what the dog did.

Yawning does appear under stress, but in lab work it appeared "to a lesser extent" than other signs. Dogs also catch yawns from people. In the first study, 21 of 29 dogs yawned after watching a stranger yawn. In another, dogs yawned more for their owner than for a stranger, with no rise in heart rate. Two other attempts found yawning in far fewer dogs, so whether dogs catch yawns is still debated.

Researchers put a life-size robot dog in a public place and watched how 492 passing dogs approached it. When the robot had a long tail, whether it was wagging changed how dogs approached. When it had a short tail, dogs approached the same way whether it was wagging or still, which suggests a short or docked tail carries less information. The effect showed mainly in dogs bigger than the robot. The authors are clear that their results do not show that docked dogs get into more fights.

Faces and coats have been studied less. A review of facial pain signs notes that skin folds on flat-faced breeds and dark coats over the brow can hide the facial movements vets use to score pain, and that some breeds lack certain facial muscles altogether. Another review notes that long or dense coats hide raised hackles. Curled tails carried over the back have not been studied.

A 2020 review by behavior clinicians found that between 28% and 82% of behavior referral caseloads involve a painful condition, with most clinics at 60 to 80%. Older estimates were 5 to 12%. The behavior problems most often linked to pain are aggression in a dog whose mood changes easily, poor impulse control, and noise sensitivity. The authors recommend investigating suspected pain first, as a precaution, rather than only after behavior work fails. The signs in the table above come from this review, apart from the facial signs.

Vets score acute pain from posture, demeanor, response to touch, attention to a wound and vocalizing. A review of pain in dogs’ faces describes squinting, flattened ears and lips drawn back. Several of these overlap with signs of fear and stress, which is why a sudden change needs a vet first.

Three of the studies cited here had support from a pet-food company, and each is marked in the sources list. The study of 60 research dogs, which found that observers’ overall impression beat any single sign, was funded by a pet-food company, and every author was its employee. The review concluding that no single behavior reliably shows a happy dog was written by two employees of a pet-food company. The second guilty-look study received a donated camera and treats from a pet-food company.

The main point of this piece, that several signals read together tell you more than any one signal, is also supported by studies with no industry funding.