How to Read Dog Body Language: Signals, Sounds and What They Really Mean

In brief

Dogs communicate multimodally—through sounds, body language, facial expressions and scent, all at the same time. The most important rule for everyday life is therefore: Never interpret a single signal in isolation. A wagging tail doesn’t automatically mean happy, and a yawn doesn’t automatically mean tired. Only the interplay of several signals, plus the situation, allows a reliable assessment.

Much of this is well supported—context-specific barking and growling, for example, or the fact that dogs wag with a bias to one side and that other dogs respond to it. Other things are not: which wagging direction means what is still open, and the popular “calming signals” are useful as a practitioners’ concept but have so far been tested scientifically only in early stages—one targeted test did not confirm the common interpretation. This article separates the two.

Two things matter more in everyday life than any list of signals. First, the sequence: a dog that works itself up into increasingly obvious signs over several minutes is telling you something different from one that shows a single signal and then loosens up again. Second, what we ourselves send—a considerable share of misunderstandings arises not in reading but in sending.

👉 The underlying principle: Behavior does not equal emotion

1. Sounds and what they mean

Dogs have a wide vocal range: barking, growling, whining, yelping, howling, grunting. These sounds are not random—they carry context-dependent information.

1.1 Barking

The most versatile signal. Its acoustic parameters—pitch, rhythm, intensity—vary with the situation, and so clearly that human listeners can classify the context of a bark better than chance: play, being left alone, a stranger (Pongrácz et al., 2005, 2006; Yin & McCowan, 2004).

As a rough guide: high-pitched, rapid barking with loose body language tends to come with excitement; deep, forceful barking tends to come with alertness; repeated barking with eye contact is often a request.

1.2 Growling

The most commonly misunderstood sound in everyday life. Growling is a communication signal, not misbehavior.

Faragó and colleagues (2010) showed that growls differ acoustically depending on context—a play growl sounds different from a growl while guarding a resource or one directed at a stranger—and that dogs recognize these differences.

A growling dog is indicating a boundary. If you punish growling, you risk the dog suppressing the warning in future—escalation then becomes harder to see coming.

1.3 Whining, yelping and howling

These occur during strong emotional arousal—pain, fear, frustration—or as contact calls in very social dogs. Nordic breeds are known for their melodic howling.

1.4 And one note that doesn’t belong in the repertoire

In short-headed breeds such as pugs or bulldogs, grunting, snoring and snorting are often anatomical in origin and a sign of labored breathing—not an expression of well-being.

This is not a side note: these noises are routinely interpreted as cute or contented, when in fact they can point to brachycephalic obstructive airway syndrome. Frequent or loud breathing noises should be checked by a veterinarian.

1.5 On breed differences

They are real: beagles and dachshunds were selected for loud hunting barks, basenjis are considered to bark little, huskies to be especially vocal. Such differences have a heritable component (MacLean et al., 2019).

Even so, every dog develops its own style—and the variation within a breed is usually greater than the variation between breeds.

One distinction belongs here, because otherwise it is easily overgeneralized: these differences mainly concern the tendency to show a signal—how quickly, how often, how loudly. There is no evidence for a breed-specific body language in which the same signals mean something different. A beagle is more likely to bark than a basenji, but a play bow looks the same in both and means the same thing.

What actually changes the readability of signals by breed is something else: anatomy. A docked tail, floppy ears, a dense coat or a short muzzle limit what is visible at all—regardless of what the dog is trying to communicate at that moment. That is the subject of section 10, and in everyday life it is the more important difference in practice.

👉 Why growling matters · When your dog barks at visitors · Breed and behavior: predisposition versus individual variation

2. Body language: posture and movement

Body language often tells you more about a dog’s state than its sounds do. It includes posture, movement patterns, muscle tension, and the position of the head, ears and tail.

A relaxed dog looks soft and loose: fluid movements, relaxed muscles, a loosely carried tail.

A tense dog shows a stiff posture, rigid or jerky movements, increased body tension—often an expression of insecurity or inner conflict.

Context decides. Without context, a single signal cannot be interpreted reliably, because the same movement pattern can occur in very different states.

👉 How can you tell your dog feels good? · Approach–avoidance conflict in dogs

3. Tail and ears—with a surprising finding

As a rough guide, always within the overall picture: loose wagging tends to be positive; a stiff, high-carried tail tends to mean high arousal; a tucked tail tends to mean fear or insecurity. In dogs with a very short tail this channel is limited—then other signals count all the more.

One finding shows particularly clearly how little “wagging tail equals joy” holds up: dogs wag asymmetrically.

The tail swung more to the right when the dogs saw something positive—their own person—and more to the left with something negative, an unfamiliar dog that appeared threatening (Quaranta et al., 2007). The authors attribute this to the different involvement of the two brain hemispheres.

And other dogs pick up on it: they responded to left-biased wagging with more tension and a higher heart rate (Siniscalchi et al., 2013).

In everyday life, however, this has little practical significance—the asymmetry can hardly be detected reliably with the naked eye. The value of the finding lies elsewhere: it shows that the same movement can accompany opposite states.

And a second limitation belongs with it, one that is almost always missing from accounts of the topic. An independent research group examined the same question using a remote-controlled, life-size dog replica and analyzed 452 encounters with free-roaming dogs. The result pointed in the opposite direction: when the replica’s tail wagged to the left, dogs were more likely to continue their approach without stopping; right-wagging more often led them to stop (Artelle, Dumoulin & Reimchen, 2011).

What both approaches support is that dogs process wagging asymmetry at all—otherwise the direction would have made no difference in either setup. Which direction signals which state, on the other hand, remains open. For practice, this only reinforces the point made above: at present, no prediction can be drawn from the wagging direction of an unfamiliar dog.

Ear position: ears laid back are associated with insecurity, pricked ears with alertness. Here too the context rule applies—and ear shape has a considerable effect on how readable they are.

👉 Laterality in dogs

4. Play signals: communication about communication

An area that is usually missing from signal overviews, even though it is one of the most interesting.

Play between dogs often looks like the real thing: chasing, body-slamming, nipping, wrestling over resources. To keep this from tipping into a genuine conflict, a level above it is needed—a signal that says: What comes next isn’t meant seriously.

4.1 The play bow: punctuation or restart?

That is exactly what the play bow does: front end low, hindquarters up, tail usually moving. In ethology it is interpreted as a metasignal—a signal about the meaning of other signals. This interpretation is widespread and plausible; what has been demonstrated, to begin with, is something narrower.

Bekoff was the first to test it. He analyzed film footage of playing adult domestic dogs as well as dog, wolf and coyote pups, and examined exactly when the bows occurred (Bekoff, 1995). They did not appear at regular intervals or every few seconds, but clustered at particular points: at the start of a sequence and around actions that are easily misunderstood—above all biting accompanied by rapid side-to-side head shaking, a movement that also occurs in aggression and in killing prey.

This uneven distribution is precisely what was actually measured. A signal that was randomly distributed could clarify nothing. One that clusters exactly where misunderstandings threaten behaves like punctuation.

The difference between the two matters more than it sounds. The finding is the distribution: bows cluster at ambiguous points. The interpretation is the conclusion drawn from it, that the bow frames the following action as play. The interpretation fits the finding well, but it goes beyond it—because whether the other dog actually reads the bow that way has not been shown. This very gap between “occurs systematically” and “is understood” runs through the whole field of dog communication.

How far this goes: The study is based on ten free-roaming adult dogs plus four pups each of dog, wolf and coyote; according to the author, the pup observations were not independent of one another. That is not much. The interpretation as a metasignal is therefore well founded but not widely replicated—and in adult dogs, a later analysis produced a different picture.

What adult dogs show. Byosiere, Espinosa and Smuts (2016) analyzed 414 play bows by 16 adult pet dogs from 229 dyadic play sequences and coded the behavior of both dogs immediately before and after each bow. Biting with head shaking did not occur around the bows, and the bowing dogs showed no more offensive behavior than their partners. Instead, bows mostly followed a brief pause; afterward both dogs were more active, and often the bowing dog ran off while the other gave chase. In 409 of the 414 cases, each dog was within the other’s field of view. The authors conclude that in adult dogs’ play the bow mainly serves to resume play that has been interrupted, and that the clarifying role described by Bekoff may matter mainly in young animals.

Here, too, the limits belong in the picture: four pairs supplied about three-quarters of the analyzed bows, a single dog 45 percent on its own. For everyday life this means: in adult dogs, the bow mainly marks the transition back into play. Whether the other dog reads it as a clarification has not been shown by either study.

4.2 The play face

Besides the play bow, there is a second play signal, and in everyday life it is the more useful one: the play face.

This means a relaxed, open mouth with soft lips, a soft gaze and loosely held facial muscles—often accompanied by a huffing, staccato pant that has nothing to do with exertion.

The key difference from the play bow: the bow is momentary, marking beginnings and transitions. The play face runs continuously alongside—it is an ongoing signal throughout the entire interaction.

That makes it a very helpful early indicator in practice. When the play face disappears and the facial muscles freeze—closed mouth, tense lips, fixed stare—the play is at risk of tipping over. And often before anything changes in the movement pattern: the two dogs are still running and bumping, but the expression is already gone.

If you watch the faces rather than the bodies during dog-dog contact, you see the shift much earlier.

4.3 Role reversal and self-handicapping—with an important caveat

Two further features are considered hallmarks of genuine play, and both are regularly presented in advice literature more strictly than the data allow.

Role reversal. The one who was just chasing gets chased. Self-handicapping. The stronger or faster dog holds back—lets itself be caught, flops down, bites much more gently than it could.

Both exist. What isn’t true is the common rule that good play has to be balanced. Bauer and Smuts analyzed the play of 55 dog pairs and found the opposite of a 50:50 pattern: only 5.4 percent of the pairs played in a roughly balanced way, and just under 22 percent completely one-sidedly—one dog in one role, the other consistently in the other (Bauer & Smuts, 2007).

And a second finding of the same study turns a common expectation on its head: older and higher-ranking dogs showed less self-handicapping, not more. The idea that the superior dog handicaps itself out of a kind of fairness does not hold up in this form.

In practice, this means: asymmetry alone is no alarm sign. Many well-functioning play pairs are asymmetrical, and a dog that mostly chases while the other mostly flees need not have a problem. What matters is something else—whether the dog in the weaker role comes back of its own accord, whether there are pauses, and whether the play face stays. The distribution of roles is the weaker information; the ability to interrupt is the stronger.

Exaggerated, loose movements. Play looks “bouncy” and inefficient, where the real thing looks purposeful and taut.

Pauses. Both dogs break off on their own, shake themselves off, sniff briefly, carry on.

4.4 How to tell it’s tipping over

This is the most important practical question in any dog-dog contact, and the signs are fairly clear.

The play face disappears—usually first. Role reversals and pauses stop. Movements become tighter and more economical. One dog repeatedly tries to get away and isn’t allowed to. Arousal rises without coming back down. And the play bow no longer appears.

One point that is often misunderstood: growling during play is normal and usually higher-pitched than serious growling. What stands out is not the growling itself, but when it gets deeper and the body language freezes at the same time.

👉 The neurobiology of play in dogs · Leash reactivity in dogs

5. Facial expressions: the subtle signals

The face also provides clues: gentle, soft eyes point to well-being; fixed or wide-open eyes showing the whites point to tension. Lip licking and looking away are often interpreted as signs of stress.

Conversely, dogs read us as well. Albuquerque and colleagues (2016) showed that dogs integrate emotional cues across sensory channels: dogs matched corresponding facial expressions and tones of voice—for dog faces and human faces.

Important for context: this demonstrates fine-grained perception and matching of emotional signals. It does not demonstrate that dogs understand our feelings in the human sense.

👉 How dogs perceive our emotions · Emotions in dogs

6. Scent: the invisible level

Through their sense of smell, dogs communicate chemically. That they take in rich information while sniffing is well established—being allowed to sniff is genuine enrichment, not dawdling.

It is also established that emotional chemosignals are transmitted: the body odor of fearful people can trigger more stress behavior in dogs (D'Aniello et al., 2018), and dogs can distinguish the odor of people under psychological stress from their baseline odor (Wilson et al., 2022).

Be careful with more far-reaching claims. That scent marks provide precise information about sex, age, health and mood is partly plausible and partly speculative. The data here are thinner than popular accounts suggest—“the dog reads the other dog’s business card” is a nice metaphor, not a finding.

That dogs take in information through odors and respond to it is established. What significance individual pieces of scent information have for the individual dog, on the other hand, has not yet been fully clarified.

👉 How olfactory processing drives canine behavior · Mental stimulation and scent games · Mantrailing

7. Calming signals in dogs: what is actually supported

The term comes from training practice, not from research. It refers to subtle behaviors such as yawning, lip licking, looking away, slow movements or stepping between others, which are interpreted as tension-reducing signals.

An honest assessment is worthwhile here, because in many training circles the concept is treated almost as established knowledge.

And it starts with the term. Behaviors such as lip licking, blinking, yawning or brief sniffing are called displacement behaviors in behavioral biology. They arise from the conflict between two opposing behavioral systems—approach versus avoidance, for example—and are classically interpreted as a release of inner tension. That makes them, first of all, an indication of an inner conflict, not automatically a message to someone else.

That is exactly the difference from the term “calming signal”: it claims that the behavior is addressed to someone—that the dog is directing something at another individual in order to reduce tension. Whether that is true is an empirical question. And it has been tested.

What speaks for the calming interpretation: A pilot study of encounters between dogs found that such behaviors occur frequently and that, when they followed an aggressive action by the other dog, a de-escalation mostly came next—though with only 73 cases and no way of separating cause from mere co-occurrence (Mariti et al., 2017).

What speaks against it: A later study of 53 dogs tested exactly this question—its title asks it explicitly: do displacement behaviors have an appeasement function? Two unfamiliar people each approached either in a mildly threatening or a neutral manner. Blinking, nose licking and lip wiping did not occur more often in the threatening trial. They occurred more often in dogs that did not bark or lunge on the leash—and equally in both trials, that is, independently of how threatening the situation was (Pedretti et al., 2023). Only head turning was more frequent in the threatening trial among the non-reactive dogs.

What remains open: The calming interpretation is ultimately a statement about the receiver, and data on that have only recently become available. In a video study, 54 dogs blinked more often when they saw other dogs blinking than when those dogs licked their noses. The authors read this as possible mimicry and as an indication that blinking could play a role in dog communication (Canori et al., 2025). This does not demonstrate appeasement, and nose licking did not trigger a comparable response. Moreover, the two interpretations are not mutually exclusive: in behavioral biology, displacement behaviors are also considered one possible origin of communication signals (Pedretti et al., 2023).

What follows from this? Not that you should ignore these behaviors. Rather, a shift in interpretation: less “he’s asking for distance,” more “he’s under tension”—without this already saying whether, or to whom, it is directed.

The practical core therefore remains correct and important: take such quiet signs seriously instead of overlooking them as cute, and defuse situations early. What is not supported is the idea that you are reading a directed message with a fixed content.

👉 Warning signals in dogs · Arousal regulation in dogs · Canine body language: what the evidence actually shows

8. The sequence matters more than the single signal

From all of the above follows something that helps more in everyday life than any list of signals: Look at the order, not the individual piece.

In training practice, the image of a ladder has become established for this. According to this image, a dog that feels uncomfortable works its way up from the bottom: first very quiet signs such as blinking, turning the head slightly away, licking the lips. Then more obvious ones such as turning the body away, freezing, lowering the body. Only after that come the loud ones: growling, showing teeth, snapping without contact—and finally the bite.

Two things make this image useful, and one caveat belongs with it.

What is useful: If you step in on the lower rungs—creating distance, defusing the situation—you can often avoid conflict altogether. And if you punish the lower rungs, you risk the dog no longer showing them. That is one possible explanation for some bites “without warning”: the warning was there and may have been trained away. This sequence has not been studied directly in dogs, however—the risk still speaks clearly against punishing early signals.

The caveat: This ladder is a practical concept, not a validated scale. Real dogs don’t work through it neatly rung by rung—some skip several, others stay at the bottom for a long time, and the order varies between individuals. It works as an aid to attention, not as a measuring instrument.

What counts in any case: a dog that keeps moving further “up” over several minutes is telling you something different from one that shows a single signal and then loosens up again.

👉 When a dog bites · The fallout of punishment · Measuring dog behavior

9. And what we send ourselves

Communication runs in both directions. A considerable share of misunderstandings arises not because we misread the dog, but because we ourselves send signals that come across differently than intended.

Approaching head-on is a clear statement among dogs. Approaching from the side in a slight curve is the polite equivalent—and with insecure dogs it regularly works where a direct approach fails.

Leaning over the dog comes across as crowding, especially from the front. Crouching down with your upper body turned away is the pressure-relieving alternative.

Prolonged, direct eye contact can act as a threat signal among dogs. That dogs seek out and tolerate eye contact with us is a special feature of the human-dog relationship—but it does not apply to unfamiliar dogs or to tense situations.

Hugging restricts freedom of movement. A research group led by Walsh analyzed the eighty most-viewed online videos of people hugging their dog and found a high rate of stress and avoidance behavior; in about two-thirds of the cases the dog used its mouth during the hug (Walsh et al., 2024).

How far this goes: The authors themselves describe their three sub-studies as pilot studies. What was analyzed were the most popular videos—a deliberately non-representative selection in which more extreme scenes are likely to be overrepresented. The figure therefore says nothing about how many dogs in general find hugs unpleasant. It says something else that is at least as important for everyday life: in exactly those scenes that are shared en masse and commented on as cute, the dog’s avoidance behavior is clearly visible to experts—and apparently goes unnoticed by the people involved.

A correction belongs here, because it is repeated wrongly almost everywhere: the widely cited “study of 250 photos” on hugging is not a study. It is a column in a popular magazine in which 250 photos found through image search were counted; the author himself explicitly did not describe it as a formal study. The solid work on the topic is the video analysis, not the photo count.

Reaching over the head and patting is often tolerated rather than enjoyed. Calm strokes on the chest, shoulder or flank are the more reliable choice.

The common denominator: much of what looks friendly to us can come across to a dog as taking up its space. That is no reason to avoid all contact—only a reason to apply the litmus test: does the dog come back of its own accord when you stop for a moment?

👉 Building trust with your dog · Safety and respect: children and dogs · How dogs decode human gestures and facial expressions

10. When channels are missing

Not every dog can send every signal, and not every dog can receive every signal. That changes how you read them considerably.

Docked or very short tail. An entire channel is lost or becomes illegible—for us and for other dogs. This can contribute to such dogs being misunderstood more easily in encounters.

Floppy ears and a dense coat. Ear position and body tension are harder to make out. In heavily coated breeds, the same goes for muscle tension and raised hackles.

Short-headed breeds. Facial expression is restricted by the structure of the face, showing teeth often looks visible anyway, and panting can hardly be used as a sign of stress, because it often occurs all the time.

Impaired sight or hearing. A deaf dog doesn’t get notice of approaches and may accordingly be startled more often; a blind dog doesn’t read body language. Both need more advance warning, not more leniency.

In practice this means: the fewer channels available, the more weight the remaining ones carry—and the more important the context becomes. With a dog that has a stubby tail and floppy ears, what remains are mainly body tension, quality of movement, gaze and breathing.

👉 Merle in dogs · Temperament in dogs

11. Emotions: present, but only indirectly readable

There is much to suggest that dogs experience emotional states and express them through body, sounds and behavior.

But one basic principle runs through all of modern behavioral research: Visible behavior is not the same as the inner feeling. We infer states from behavior—we have no direct access to experience.

The same behavior can have different causes, and similar states can show up differently. That is why restraint with clear-cut attributions is called for—and careful, multidimensional observation matters more than quick labels.

Two dimensions help with sorting: arousal—how intense—and valence—pleasant or unpleasant. The two vary independently of each other, and neither can be derived from a single signal.

👉 Comforting a fearful dog · Reactivity in dogs · Frustration in dogs

12. Neurobiology: a young field

Imaging in awake dogs is a young line of research, and the body of findings is correspondingly narrow.

What is established: the dog brain has voice-sensitive areas and processes the emotional tone of sounds in a way comparable to the human brain (Andics et al., 2014). And the reward system responds to social attention such as praise (Cook et al., 2016).

What is not established are the more far-reaching claims you often read—for example, that the amygdala and prefrontal cortex specifically control communication. Such attributions come from research on other species.

And one more point that applies across the board: these methods measure blood-flow signals, not neurotransmitters or feelings.

👉 The neurology of dog behavior

13. What the research does not show

• Single signals are ambiguous. Wagging, yawning or licking have no fixed meaning. Only the overall context is reliable.
• The calming function is not confirmed. A pilot study supports the concept, a targeted test with 53 dogs did not find the expected pattern; data on how receivers respond exist so far only for blinking.
• Perceiving emotions is not understanding emotions. Dogs match emotional signals with fine precision; that does not imply that they understand our feelings.
• Scent communication is only partly decoded. Emotional chemosignals are established; precise scent profiles are conjecture.
• Brain imaging in dogs is young. Findings exist mainly on voice and reward, from small samples; detailed attributions to brain areas are provisional.
• The escalation ladder is a practical concept, not a validated scale. Real dogs don’t necessarily go through the rungs in order; its value lies in drawing attention to the sequence.
• Play signals are well described but not quantified. That the play bow, role reversal and self-handicapping characterize play is established observation—what function the bow serves is judged differently by the two available analyses, and the point at which play tips over has not been fixed in thresholds.
• Breed differences are measurable but weak predictors. The variation within a breed is usually greater than the variation between breeds.

👉 A closer look at barking and growling: Vocal communication in dogs

14. Conclusion

The language of dogs is multilayered—and readable if you look at signals in combination and in context instead of over-interpreting them one at a time.

If you learn this, you avoid misunderstandings, recognize conflicts earlier and strengthen your relationship. And if you stay honest about what is not yet settled, you read more accurately than someone who has a translation ready for every signal.

The most useful rule remains the simplest: one signal is not a sentence. It takes several together to make one.

Key takeaways on reading dog body language

Never interpret a single signal in isolation. Dogs send through sounds, body, face and scent at the same time. Only the interplay plus the situation allows an assessment.

Barking and growling carry context information. People can classify the context of a bark better than chance (Pongrácz et al., 2005, 2006; Yin & McCowan, 2004), and growls sound different depending on the situation—dogs hear the difference (Faragó et al., 2010).

Growling is a warning, not misbehavior. If you punish it, you risk the warning disappearing while the conflict remains.

The play bow is momentary, not a continuous signal. In the older study it clustered around actions that are easily misunderstood (Bekoff, 1995); in adult pet dogs it mainly followed pauses in play (Byosiere et al., 2016). The signal that runs continuously is the play face—and its disappearance is often one of the earliest signs that play is tipping over.

Good play doesn’t have to be balanced. Only 5.4 percent of 55 dog pairs played in a roughly symmetrical way, and older and higher-ranking dogs held themselves back less, not more (Bauer & Smuts, 2007). What matters is whether play can be interrupted, not how the roles are distributed.

The calming function is not confirmed. In a test with 53 dogs, blinking, nose licking and lip wiping did not occur more often in threatening situations (Pedretti et al., 2023). Take them seriously, yes—read them as a directed message with a fixed content, no.

The sequence matters more than the single signal. If you step in at the quiet rungs, you can often avoid conflict altogether.

We send signals too. Approaching head-on, leaning over the dog, prolonged eye contact and hugging can come across to dogs as taking up their space; for hugging, a video analysis supports this (Walsh et al., 2024). The litmus test: does the dog come back of its own accord when you stop for a moment?

👉 Emotional contagion: can human stress affect dogs?

References

Albuquerque, N., Guo, K., Wilkinson, A., Savalli, C., Otta, E., & Mills, D. (2016). Dogs recognize dog and human emotions. Biology Letters, 12(1), 20150883. https://doi.org/10.1098/rsbl.2015.0883

Andics, A., Gácsi, M., Faragó, T., Kis, A., & Miklósi, Á. (2014). Voice-sensitive regions in the dog and human brain revealed by comparative fMRI. Current Biology, 24(5), 574–578. https://doi.org/10.1016/j.cub.2014.01.058

Artelle, K. A., Dumoulin, L. K., & Reimchen, T. E. (2011). Behavioural responses of dogs to asymmetrical tail wagging of a robotic dog replica. Laterality, 16(2), 129–135. https://doi.org/10.1080/13576500903386700

Bauer, E. B., & Smuts, B. B. (2007). Cooperation and competition during dyadic play in domestic dogs, Canis familiaris. Animal Behaviour, 73(3), 489–499. https://doi.org/10.1016/j.anbehav.2006.09.006

Bekoff, M. (1995). Play signals as punctuation: The structure of social play in canids. Behaviour, 132(5–6), 419–429. https://doi.org/10.1163/156853995X00649

Byosiere, S.-E., Espinosa, J., & Smuts, B. (2016). Investigating the function of play bows in adult pet dogs (Canis lupus familiaris). Behavioural Processes, 125, 106–113. https://doi.org/10.1016/j.beproc.2016.02.007

Canori, C., Travain, T., Pedretti, G., Fontani, R., & Valsecchi, P. (2025). If you blink at me, I'll blink back. Domestic dogs' feedback to conspecific visual cues. Royal Society Open Science, 12(2), 241703. https://doi.org/10.1098/rsos.241703

Cook, P. F., Prichard, A., Spivak, M., & Berns, G. S. (2016). Awake canine fMRI predicts dogs' preference for praise vs. food. Social Cognitive and Affective Neuroscience, 11(12), 1853–1862. https://doi.org/10.1093/scan/nsw102

D'Aniello, B., Semin, G. R., Alterisio, A., Aria, M., & Scandurra, A. (2018). Interspecies transmission of emotional information via chemosignals: From humans to dogs. Animal Cognition, 21(1), 67–78. https://doi.org/10.1007/s10071-017-1139-x

Faragó, T., Pongrácz, P., Range, F., Virányi, Z., & Miklósi, Á. (2010). “The bone is mine”: Affective and referential aspects of dog growls. Animal Behaviour, 79(4), 917–925. https://doi.org/10.1016/j.anbehav.2010.01.005

MacLean, E. L., Snyder-Mackler, N., vonHoldt, B. M., & Serpell, J. A. (2019). Highly heritable and functionally relevant breed differences in dog behaviour. Proceedings of the Royal Society B, 286(1912), 20190716. https://doi.org/10.1098/rspb.2019.0716

Mariti, C., Falaschi, C., Zilocchi, M., Fatjó, J., Sighieri, C., Ogi, A., & Gazzano, A. (2017). Analysis of the intraspecific visual communication in the domestic dog: A pilot study on the case of calming signals. Journal of Veterinary Behavior, 18, 49–55. https://doi.org/10.1016/j.jveb.2016.12.009

Pedretti, G., Canori, C., Biffi, E., Marshall-Pescini, S., & Valsecchi, P. (2023). Appeasement function of displacement behaviours? Dogs' behavioural displays exhibited towards threatening and neutral humans. Animal Cognition, 26(3), 943–952. https://doi.org/10.1007/s10071-023-01742-9

Pongrácz, P., Molnár, C., Miklósi, Á., & Csányi, V. (2005). Human listeners are able to classify dog barks recorded in different situations. Journal of Comparative Psychology, 119(2), 136–144. https://doi.org/10.1037/0735-7036.119.2.136

Pongrácz, P., Molnár, C., & Miklósi, Á. (2006). Acoustic parameters of dog barks carry emotional information for humans. Applied Animal Behaviour Science, 100(3–4), 228–240. https://doi.org/10.1016/j.applanim.2005.12.004

Quaranta, A., Siniscalchi, M., & Vallortigara, G. (2007). Asymmetric tail-wagging responses by dogs to different emotive stimuli. Current Biology, 17(6), R199–R201. https://doi.org/10.1016/j.cub.2007.02.008

Shepherd, K. (2009). The canine ladder of aggression. In D. F. Horwitz & D. S. Mills (Eds.), BSAVA manual of canine and feline behavioural medicine (2nd ed., pp. 13–16). British Small Animal Veterinary Association.

Siniscalchi, M., Lusito, R., Vallortigara, G., & Quaranta, A. (2013). Seeing left- or right-asymmetric tail wagging produces different emotional responses in dogs. Current Biology, 23(22), 2279–2282. https://doi.org/10.1016/j.cub.2013.09.027

Wilson, C., Campbell, K., Petzel, Z., & Reeve, C. (2022). Dogs can discriminate between human baseline and psychological stress condition odours. PLOS ONE, 17(9), e0274143. https://doi.org/10.1371/journal.pone.0274143

Walsh, E. A., Meers, L. L., Samuels, W. E., Boonen, D., Claus, A., Duarte-Gan, C., Stevens, V., Contalbrigo, L., & Normando, S. (2024). Human-dog communication: How body language and non-verbal cues are key to clarity in dog directed play, petting and hugging behaviour by humans. Applied Animal Behaviour Science, 272, 106206. https://doi.org/10.1016/j.applanim.2024.106206

Yin, S., & McCowan, B. (2004). Barking in domestic dogs: Context specificity and individual identification. Animal Behaviour, 68(2), 343–355. https://doi.org/10.1016/j.anbehav.2003.10.016

Notes on the sources

The evidence on wagging direction is not consistent: laboratory studies using videos and silhouettes (Siniscalchi et al., 2013) and field studies using a dog replica (Artelle et al., 2011) arrive at opposite directions. What both have in common is that the asymmetry is processed; its meaning remains open. This is covered in detail in the research article on laterality.

Two analyses are available on the function of the play bow. Bekoff (1995) draws on ten adult dogs plus four pups each of dog, wolf and coyote and found bows clustered around easily misunderstood actions; the interpretation as a metasignal is therefore well founded but not widely replicated. Byosiere et al. (2016) analyzed 414 bows by 16 adult pet dogs, found no support for this in adult dogs, and interpret the bow there mainly as a resumption of play after a pause; four pairs supplied about three-quarters of the data. Whether the partner understands the bow as a clarification was not tested by either study.

The figures from Bauer and Smuts (2007) concern 55 dog pairs: 5.4 percent roughly symmetrical play, around 22 percent completely asymmetrical, less self-handicapping in older and higher-ranking animals. Advice literature often cites the study as evidence that role reversal and self-handicapping are the norm. That reverses its finding: it shows how rare balanced play actually is.

On Walsh et al. (2024): these are three pilot studies of the most popular online videos—30 on play, 80 each on hugging and petting. The selection is deliberately non-representative; the results describe what can be seen in particularly successful videos, not the distribution in dogs’ everyday lives. The widely cited count of 250 photos on hugging, by contrast, is not a scientific study but a magazine piece, and is not used as evidence here.

On calming signals, the evidence should be assessed considerably more cautiously than common accounts suggest. The term comes from the training literature, not from research. Mariti et al. (2017) is a study explicitly described as a pilot study. Pedretti et al. (2023) tested the question specifically with 53 dogs and found that blinking, nose licking and lip wiping were not related to how threatening the situation was, but occurred more often in dogs that did not bark or lunge on the leash—regardless of the test condition. Head turning was the only behavior that occurred more frequently in the threatening trial among non-reactive dogs. The claim that these signals serve to reduce tension in the other individual is therefore not confirmed, but not refuted either; data on the receiver side exist so far only for blinking, and there they showed possible mimicry, not appeasement (Canori et al., 2025). The practical recommendation to take them seriously as signs of stress is unaffected by this.

The statements on wagging asymmetry draw on Quaranta et al. (2007) with 30 dogs and Siniscalchi et al. (2013). It is established that the asymmetry occurs and that other dogs respond to it; which direction means what remains open because of the opposite finding by Artelle et al. (2011). In everyday life, the asymmetry cannot be reliably detected by eye anyway.

Albuquerque et al. (2016) demonstrates the cross-modal matching of emotional signals. It does not show that dogs understand human emotional states; the study works with preferential looking-time measures.

The imaging findings (Andics et al., 2014; Cook et al., 2016) are based on small samples of specially trained dogs and record blood-flow signals as a correlate of neural activity.

The note on breathing noises in short-headed breeds reflects the veterinary consensus on brachycephalic obstructive airway syndrome and is included here as a safety note.

The description of the other play signals—role reversal, self-handicapping, loose movements, pauses—reflects established ethological observation and, apart from Bauer and Smuts (2007), is not individually cited. Criteria for the point at which play turns serious are descriptive and have not been validated as thresholds.

The sequence from quiet to obvious signals goes back to the British veterinarian and behaviorist Kendal Shepherd, who presented it as the “canine ladder of aggression” in the BSAVA manual of behavioral medicine. It was conceived as a teaching model, not as a validated scale; as far as our research shows, no empirical test of the order of the rungs exists. It is well founded in learning psychology that punishment can suppress a visible behavior without changing the underlying state; the specific sequence from punished early signal to a bite without warning, however, has not been studied directly in dogs.

On the choice of term for displacement behaviors: the term refers to behaviors that arise from the conflict between opposing behavioral systems and are classically interpreted as a release for the individual. Compared with “calming signal,” it is the more cautious term, because it does not presuppose that the behavior is addressed to someone else. Pedretti et al. (2023) use it in the title of their study and examine the appeasement function as an open question; at the same time they point out that displacement behaviors are a possible origin of communication signals, so the two terms are not mutually exclusive.

The play face—a relaxed, open mouth with soft lips—has long been described as an expressive behavior in comparative behavioral research. The statement made in the text that it is often one of the earliest indicators of play tipping over reflects practical experience and has not been quantified in thresholds.

The statements about human signaling are based on ethological observation and practical experience. The statement on hugging draws on a peer-reviewed video analysis of self-selected online media without physiological measurement.

The statements on breed differences follow MacLean et al. (2019). The heritability determined there concerns differences between breeds; the predictive power of breed for the individual dog is considerably lower and is covered in detail in the research article on breed and behavior.

Frequently asked questions about reading dog body language

How to get better at understanding what your dog is expressing