Canine Body Language: What the Evidence Actually Shows

1. Reading Behavior, Not Minds

1.1 What Body Language Is Taken to Mean

Almost every account of canine body language promises a translation. A lowered tail means fear, a yawn means stress, a play bow means an invitation, a lip lick means an attempt to calm the other party. Owners and trainers use these equivalences as a vocabulary.

The research literature is more useful than that and considerably less tidy. What has been measured is that particular movements of the tail, face and body occur more often in some situations than in others, that other dogs and people respond to some of them, and that some of them are associated with physiological measures of arousal. What has rarely been measured is the step from a movement to the state or intention behind it, and that step is where most popular interpretation lives (why an observed behavior is not yet an identified emotion).

1.2 How to Read the Evidence

Three kinds of study run through this article, and they answer different questions. Production studies record what dogs do in defined situations and ask whether a behavior is more frequent in one context than another. Reception studies expose dogs or people to a signal and record the response. Interpretation studies ask human observers to judge what a dog is feeling and compare the answer with the situation the dog was in.

A behavior can pass the first test without passing the second, so each section below says which kind of evidence carries the claim (how a behavioral construct is defined before it is measured).

2. The Tail

2.1 Asymmetric Wagging

The tail is the signal owners watch most, and its best-known research finding concerns direction rather than speed or height. In a study of 30 mixed-breed dogs placed in an enclosure with cameras that tracked tail position, dogs were shown four stimuli in turn: their owner, an unfamiliar person, a cat and an unfamiliar, dominant dog (Quaranta, Siniscalchi & Vallortigara, 2007).

Wagging was biased to the right side of the dog's body when the dogs saw their owner, with weaker right-sided biases toward the unfamiliar person and the cat. When the dogs saw the unfamiliar dog, the bias shifted to the left. The authors interpreted the pattern in terms of hemispheric specialization, with approach-related responses linked to left-hemisphere activity and withdrawal-related responses to the right hemisphere.

2.2 Do Other Dogs Respond to It?

A production finding does not establish that the signal carries information to anyone. A later study from the same group tested reception directly. Dogs watched videos of another dog wagging with a left or right bias while their behavior and heart rate were recorded (Siniscalchi, Lusito, Vallortigara & Quaranta, 2013).

Dogs that watched left-biased wagging showed higher cardiac activity and more anxious behavior than dogs that watched right-biased wagging. The asymmetry may be a by-product of lateralized brain activity that other dogs have come to respond to — the difference between a cue that can be read and a signal that was produced in order to be read.

The direction of the effect has not been consistent across methods. In a field experiment by a different research group, free-ranging dogs encountered a life-size, remotely controlled dog replica whose tail wagged with a left or right bias. Across 452 interactions, a significantly greater proportion of dogs approached the model continuously without stopping when the tail wagged to the left, while right-biased wagging was more likely to be followed by a stop (Artelle, Dumoulin & Reimchen, 2011). That runs counter to what the later video study would lead one to expect. Together, the studies support the view that dogs respond to asymmetric wagging, not a fixed meaning for either direction.

2.3 What the Tail Findings Do Not License

The enclosure and video studies used controlled presentations and instrument-based measurement. Whether people can see the asymmetry reliably in real time has not been tested, and with conflicting results on direction, the studies do not provide a rule that a person could apply to a wagging dog on a walk.

2.4 Height, Speed and Stiffness

Some aspects of tail posture and movement have been studied empirically. In an experiment at off-leash parks, 492 dogs encountered a life-size robotic dog whose tail was long or short and either wagging or still. Larger dogs were less cautious and more likely to approach the long wagging tail than the long still tail, and dogs responded with a raised head and tail to the long wagging tail — but tail motion made no measurable difference when the model's tail was short (Leaver & Reimchen, 2008).

Posture has also been quantified between dogs. In one group of domestic dogs, 7 body postures and 24 behaviors were analyzed as status indicators: high posture and muzzle bite qualified best as formal dominance indicators, body tail wag was the best formal submission indicator, and lowering of posture was the best status indicator across 91 percent of dyads. Rank did not correlate with age or weight, and the hierarchy rested on submission rather than aggression (van der Borg, Schilder, Vinke & de Vries, 2015). It describes one group, not a general code.

What has not been validated are the one-to-one readings popular guides offer — a high, stiff, slowly wagging tail as a warning, a low, loose, broad wag as appeasement. Those descriptions come largely from clinical and observational tradition, and interpretations of height, speed, amplitude and stiffness remain poorly validated and highly dependent on context.

That does not make them useless. It makes them working descriptions to be checked against the individual dog, not established indicators. Breed-typical tail carriage varies so widely that a height which signals tension in one dog is the resting position of another, and docked or tightly curled tails remove much of the information the guides rely on — consistent with the finding that a short tail carried motion cues less effectively to other dogs (Leaver & Reimchen, 2008).

3. The Face

3.1 Measuring Faces Without Interpreting Them

DogFACS describes facial movements as individual action units tied to specific muscles — an inner brow raise, a lip corner pull, an ear flattening — without assigning a meaning to any of them. That separation is what allows a study to ask whether a movement is associated with a situation rather than assuming the answer in the coding scheme (Bremhorst et al., 2019) (other ways researchers infer emotional state from measurable behavior).

3.2 Anticipation Versus Frustration

Positive anticipation and frustration can occur in nearly identical situations: a dog waiting for food that is about to arrive, and the same dog waiting for food that does not. In a controlled setup that produced both conditions, Labrador retrievers were coded with DogFACS while they expected a food reward or were prevented from reaching it (Bremhorst, Sutter, Würbel, Mills & Riemer, 2019).

The "ears adductor" action was more common in the positive condition. Blinking, lips parting, jaw dropping, nose licking and ear flattening were more common in the negative condition. The authors presented this as a step toward inferring emotional states of different valence from facial expression in situations that look alike from the outside (what frustration is and how it has been measured in dogs).

3.3 Did the Result Hold?

A new sample of 28 dogs was tested with a similar design across two reward types, food and toys (Bremhorst, Mills, Würbel & Riemer, 2022). The earlier associations were replicated: ears adductor with the positive condition, and ear flattening, blinking, lips parting, jaw dropping and nose licking with the negative condition. Four further facial actions were also more common in the negative condition, and all actions except the upper lip raiser were independent of reward type.

The authors then assessed diagnostic accuracy. Ear flattening and ears downward had relatively high sensitivity but low specificity, the other negative correlates showed the opposite pattern, and ears adductor had excellent specificity but low sensitivity. Used individually, none of the facial actions would allow consistent correct classification of the associated emotion (Bremhorst et al., 2022).

A replication by the same research group under a similar protocol is not the same as independent confirmation, but it is more than most findings in this area have. It supports treating these actions as candidate indicators of anticipation and frustration in that kind of situation. It does not show that a blink or a nose lick means frustration wherever it occurs.

3.4 Faces Change With Human Attention

Facial expressions in animals were long treated as involuntary displays of emotional state. Kaminski, Hynds, Morris and Waller (2017) tested that assumption by presenting 24 dogs with a person who was either facing them or turned away, with or without food.

Dogs produced significantly more facial movements when the person was attentive. The presence of food, an arousing but non-social stimulus, did not change facial activity. The authors concluded that dog facial expressions are sensitive to the attentional state of a human audience and may be, at least in part, active attempts to communicate rather than purely automatic readouts of emotion.

Similarly, when food was withheld, blinking, nose licking and lip wiping were more likely when a human partner was visible than when no one was (Pedretti et al., 2022), as the same research group later summarized the result (Pedretti et al., 2023).

The kind of audience matters as well. When food was withheld with another dog, a person or no visible audience present, 46 dogs showed more ears downward, nose licking, lip wiping, tail wagging, whining and panting with a dog present than with no audience, kept more distance from the dog than from the person, and differed in several facial actions between the two audiences; post-test cortisol did not differ between conditions (Pedretti, Canori, Costantini, Palme, Valsecchi & Marshall-Pescini, 2024). An audience effect shows that the social partner changes what is produced, not what the dog intends to communicate.

3.5 The Inner Brow Raise

One facial movement has attracted particular attention because of its anatomy. Dissections of a small number of dog and wolf heads showed that the levator anguli oculi medialis, a muscle that raises the inner eyebrow intensely, was found as a distinct muscle in five of the six dogs examined, could not be located in a Siberian husky, and was not present as a distinct muscle in the four wolves (Kaminski, Waller, Diogo, Hartstone-Rose & Burrows, 2019). In behavioral observations, dogs produced the eyebrow movement more often and at higher intensity than wolves, and the highest-intensity movements occurred only in dogs.

The authors proposed that this movement, which makes the eyes appear larger and resembles a human expression of sadness, may elicit a nurturing response in people, and that dogs with expressive eyebrows had a selective advantage during domestication. That evolutionary account is a hypothesis built on a comparison of anatomy and behavior, and later comparative anatomy has weakened it (the wider debate about what domestication changed in dogs).

In a preliminary, qualitative dissection of ten specimens from several canid species, both the levator anguli oculi medialis and a second muscle implicated in dogs' facial communication, the retractor anguli oculi lateralis, were found in species other than dogs (Sexton, Diogo, Subiaul & Bradley, 2024). Coyotes turned out to have a well-developed levator anguli oculi medialis like dogs, in contrast to the modified or absent muscle in gray wolves, and those authors suggest that the muscle is an ancestral trait lost in gray wolves rather than a novelty produced by domestication (Cunningham, Shankar, vonHoldt, Brzeski & Kienle, 2024). In 2026, both muscles were identified in all 12 adult specimens examined from four Neotropical canid species, with conserved architecture (de Souza-Junior, Souza, Viotto-Souza, Costa, Crivelaro, Kasper & Abidu-Figueiredo, 2026). The dog–wolf difference measured in 2019 stands; the explanation that domestication created the muscles does not, though domestication may have shaped their use.

3.6 The Brow Raise and Adoption

A related study used shelter rehoming as a real-world measure of human preference. Dogs that produced the inner brow raise more frequently were rehomed more quickly (Waller, Peirce, Caeiro, Scheider, Burrows, McCune & Kaminski, 2013).

The result is an association from shelter data, not an experimental manipulation, consistent with people responding to this facial movement, whether or not they notice it.

Whether dogs produce the movement for a human audience is a separate question. When dogs expected a reward either from a person facing them or from an automated apparatus, the inner brow raise did not show the pattern a communicative function would predict; it was associated with eye movements, and the authors concluded that a lower-level explanation is possible (Bremhorst, Mills, Stolzlechner, Würbel & Riemer, 2021).

3.7 What the Face Research Does and Does Not Establish

Together, the facial studies establish that dogs' facial movements can be coded reliably, that some of them differ between situations of different valence, that facial activity rises when a person is watching, and that at least one movement has been associated with shelter rehoming speed. They do not provide a dictionary.

Most of the associations were measured in specific experimental situations, several in single breeds, and they describe changes in frequency rather than signals that are present in one state and absent in another. Short-nosed breeds with altered facial anatomy may also produce or display some of these movements differently, which none of the studies here was designed to test (how extreme facial conformation affects dogs).

4. "Calming Signals": Origin and Evidence

4.1 Where the Concept Comes From

The idea that dogs use specific behaviors to calm others and avoid conflict was popularized by the Norwegian trainer Turid Rugaas, whose book described head turning, nose licking, yawning, sniffing the ground, paw lifting, shaking off, freezing and other behaviors as "calming signals" (Rugaas, 2006). The concept spread widely in training and owner education, often presented as an established part of canine communication.

The book was based on practical observation rather than on controlled study, which is why its functional claims needed testing.

4.2 Signals Between Dogs

The first systematic test of the concept in dog–dog encounters observed 24 dogs, 12 female and 12 male, each meeting four different dogs off-leash for five minutes: a familiar and an unfamiliar dog of the same sex, and a familiar and an unfamiliar dog of the other sex (Mariti, Falaschi, Zilocchi, Fatjó, Sighieri, Ogi & Gazzano, 2017).

In total, 2,130 of the behaviors described by Rugaas were recorded, with head turning, nose licking, freezing and turning away among the most frequent. When such behaviors were shown after an aggressive action by the other dog, the aggressive display de-escalated in 79.4 percent of cases, increased in 5.5 percent and remained unchanged in 15.1 percent. The authors concluded that these behaviors may have a role in social facilitation and in preventing further aggression (how dogs manage conflict with one another).

4.3 What That Study Can and Cannot Show

The de-escalation figure rests on 73 instances in which a putative calming signal followed aggression, which is a small number from which to generalize. As later researchers pointed out, the study also grouped many different behaviors — displacement activities such as nose licking and sniffing alongside postural behaviors such as turning away or lying down — into a single category, which makes it impossible to say which of them, if any, did the calming (Pedretti et al., 2023).

Most importantly, the design is observational. A behavior followed by de-escalation is compatible with that behavior causing de-escalation, and equally compatible with both reflecting a shift in the encounter that was already under way. The study is best read as showing that the concept is worth taking seriously, not as confirming it.

4.4 Signals Toward People

Firnkes, Bartels, Bidoli and Erhard (2017) examined two of the proposed signals, looking away and lip licking, in 116 adult dogs guided by their owners through a standardized behavioral test with unfamiliar people who behaved neutrally, in a friendly way or in a threatening way.

Both behaviors were directed at the testers more often in threatening and conflict-laden situations, and the authors concluded that both may serve as appeasement signals toward humans. Two further results complicate the picture. Dogs showed significantly fewer of these behaviors when the threat was most intense, which the authors suggested may reflect a switch to other strategies such as clearly submissive behavior or escape. And lip licking was a common part of dogs' ordinary greeting behavior toward people.

4.5 A Direct Test of the Conflict Prediction

If these behaviors are appeasement signals, they should occur more in conflict than in neutral situations, and more in dogs that are not inclined to escalate. Pedretti, Canori, Biffi, Marshall-Pescini and Valsecchi (2023) tested both predictions with 53 dogs, each approached on leash by two unfamiliar women, one walking slowly with a bent posture and a fixed stare and one approaching normally and in a relaxed manner.

About half of the dogs barked or lunged at the threatening person, compared with about a fifth toward the neutral person. Blinking, nose licking and lip wiping were more frequent in dogs that did not bark or lunge — but equally so in the neutral and the threatening condition. Head turning was more frequent in non-reactive dogs only during the threatening approach, and paw lifting lasted longer in the threatening condition regardless of how the dog responded.

4.6 A Receiver Test

The appeasement account is ultimately about receivers, and receiver data have only recently begun to appear. In a video study, 54 dogs watched other dogs blinking, licking their noses or looking attentively while still. The viewing dogs blinked more during the blinking videos than during the nose-licking videos, which the authors read as a possible mimicry response and an indication that blinking may have a role in dog communication. Heart rate variability rose in all video sessions, suggesting the videos were not stressful, and blink synchrony was not examined (Canori, Travain, Pedretti, Fontani & Valsecchi, 2025).

The study shows a receiver response to one behavior on the calming-signal list. It does not show that the response amounts to appeasement, and the nose-licking videos did not produce a comparable response.

4.7 What the Studies Add Up To

The behaviors described as calming signals are real and frequently observed in dog–dog and dog–human interactions, and in the human-approach study several of them were associated with a non-reactive behavioral profile. What has not been established is the specific function the name assigns to them. In the most direct test of key predictions, most of these behaviors did not increase in the conflict situation as the appeasement account predicts, and the authors concluded that their appeasement function cannot yet be confirmed (Pedretti et al., 2023).

The same authors, drawing on an earlier critical editorial (Overall, 2017), argue that whether a behavior functions as an appeasement signal can only be judged by studying the responses of receivers and the concurrent behavior of the sender, not from its presence alone (Pedretti et al., 2023). A label that assigns a function in advance tends to be applied in every situation where the behavior appears, which is precisely what the data do not support.

4.8 Why the Label Still Matters

The practical value of the calming-signal idea was never mainly about function. It taught owners and trainers to notice small behaviors that they had previously overlooked, and that change is valuable whatever the eventual verdict on appeasement turns out to be.

5. Stress Indicators and Signals

5.1 The Same Behavior, Two Explanations

Many of the behaviors discussed as calming signals also appear in lists of stress indicators: nose licking, yawning, paw lifting, scratching, shaking off. The two accounts are not mutually exclusive. A behavior can originate as a by-product of arousal or motivational conflict and still carry information to an observer, and in ethology such displacement activities are one proposed origin of communicative signals (Pedretti et al., 2023).

5.2 Behavior and Physiology Together

The foundational canine work on stress responses measured behavior alongside salivary cortisol and heart rate while dogs were exposed to different types of stimuli, and showed that dogs mount measurable stress responses with individual differences in their course (Beerda, Schilder, van Hooff, de Vries & Mol, 1998). Studies of this kind are one foundation for treating certain behaviors as stress-related, though not the only one on which modern stress ethograms rest.

They also show why individual behaviors are weak indicators on their own. Physiological and behavioral measures do not move in lockstep, and in a controlled study of 60 dogs, cortisol, ACTH, heart rate variability, panting, whining and body shaking differed with arousal only within negative-valence scenarios (Flint, Weller, Parry-Howells, Ellerby, McKay & King, 2024). A behavior can tell an observer that something is happening without telling them what (how arousal is measured and why it is not the same as distress).

5.3 Context Changes the Reading

Lip licking illustrates the problem well. In the standardized human-approach test it was more frequent in tense situations and also a common component of friendly greeting (Firnkes et al., 2017). In the facial-coding work, nose licking was associated with frustration in a waiting task (Bremhorst et al., 2019). And in the threat test it was associated with a non-reactive stance regardless of whether the approach was threatening (Pedretti et al., 2023).

Each of these findings is plausible, and none of them licenses a single meaning for the behavior. What they support is a narrower rule: a change in the frequency of such behaviors in a particular dog, in a particular situation, is worth attending to, and its meaning has to be worked out from the situation and the rest of the dog.

6. Play Signals

6.1 The Classic Account of the Play Bow

The play bow — forequarters lowered, hindquarters raised — is the most stereotyped play behavior in dogs and appears in related canids. The influential account of its function came from Bekoff (1995), who, as later summarized, observed that play bows occurred in association with behaviors that could be misinterpreted as aggressive, particularly bite-shakes, and proposed that bows clarify playful intent. His sample included infant dogs, infant wolves and infant coyotes as well as adult dogs (Byosiere et al., 2016).

6.2 What Adult Dogs Actually Do

Byosiere, Espinosa and Smuts (2016) analyzed 414 play bows by 16 adult pet dogs during 229 bouts of dyadic play, coding the behaviors of both dogs immediately before and after each bow. Play bows most often followed a brief pause. After the bow, both dogs were more active, synchronous behaviors increased, and the bowing dog often ran away while the partner chased.

The study found no support for the clarification account in adult dogs. No bite-shakes were observed around play bows, bowers did not show more offensive behaviors than partners, and it was partners rather than bowers who became more offensive after a bow. The authors concluded that in adult dyadic play, bows function mainly to reinitiate play after a pause, and suggested that the clarification function Bekoff described may apply mainly to young animals (what play research has established in dogs).

6.3 A Visual Signal Directed at a Partner

In 409 of the 414 bows, the two dogs were within each other's visual field. Of the four remaining bows that could be assessed, the bowing dog barked in three (Byosiere et al., 2016). That pattern — signals given where they can be seen, and attention-getting behavior added where they cannot — is one of the clearer indications in this literature that a canine display is used as a signal to a specific receiver.

6.4 Limits of the Play Bow Data

Four dyads produced about three quarters of the bows analyzed, and a single dog produced 45 percent of them.

7. How Well People Read Dogs

7.1 Owners Notice the Obvious Signs

Knowing what dogs do is only half of the problem; the other half is whether people see it. In a questionnaire study of 1,190 owners, the behaviors most often identified as signs of stress were trembling and whining, followed by aggressiveness, excessive barking and panting. Subtler behaviors such as looking elsewhere, turning the head, yawning and nose licking were reported much more rarely (Mariti, Gazzano, Moore, Baragli, Chelli & Sighieri, 2012).

The authors concluded that few owners were likely to recognize stress at an early stage. Because this was a survey of what owners said rather than a test of what they saw, it describes owners' awareness of the behaviors rather than their accuracy in observing them. That is still the relevant quantity for education: a behavior an owner does not think of as meaningful is unlikely to change what the owner does.

7.2 Children Misread Faces Predictably

Children show a specific and consistent error. When 89 children aged four to seven and 30 adults interpreted photographs of dog and human facial expressions, the adults made hardly any mistakes with aggressive faces, while 67 percent of four-year-olds misread aggressive dog faces, with errors on close to 40 percent of them. The proportion fell with age, but just under half of seven-year-olds still misread aggressive dog faces, with 17 percent errors, while aggressive human faces were not misread in this way. The most frequent error was reading an aggressive dog face as happy (Meints & Racca, 2026).

Photographs remove movement and context, but the direction of the error is the one that matters for safety, and it explains why teaching children to recognize warning signals has to be concrete rather than general (why children are the most difficult case in bite prevention).

7.3 Positive States Are the Harder Problem

Most research on canine body language concerns negative or ambiguous states. No single behavioral or physiological indicator has been validated as a reliable, context-independent marker of positive emotion in dogs (Csoltova & Mehinagic, 2020). In the controlled comparison above, qualitative behavioral assessment scores were associated with both valence and arousal and were considered the best indicator of positive valence (Flint et al., 2024), a promising lead, not a finished tool. Confident statements about a dog being happy, relaxed or enjoying an interaction still rest on less measurement than statements about stress. The absence of stress behaviors is not evidence of a positive state, and the facial actions associated with positive anticipation in one experimental task (Bremhorst et al., 2019, 2022) have not been shown to indicate wellbeing in general.

8. Reading a Whole Dog in Context

8.1 Sequences, Not Single Signals

The studies reviewed here show individual behaviors changing in frequency between situations, not behaviors that identify a state on their own. The practical implication is that interpretation should rest on combinations and sequences — what the dog was doing before, what changed, what happened next — rather than on a single movement matched to a list.

8.2 The Individual Baseline

Dogs differ in how often they show these behaviors when nothing in particular is happening. Knowing a dog's ordinary behavior in familiar, low-demand situations is what makes changes interpretable (how individual dogs differ in consistent ways).

8.3 Morphology Can Remove Information

Docked tails, heavy coat around the face, drooping or cropped ears and shortened muzzles all change what an observer can see, and may change what other dogs can see as well. For tails, a short length reduced how well motion cues reached other dogs (Leaver & Reimchen, 2008); for ears, coat and muzzle shape, no comparable data were found among the sources reviewed here. In such dogs, a missing signal should carry less weight.

8.4 The Guilty Look as a Warning Case

The so-called guilty look is a useful caution about interpreting body language through a story. In the study that tested it, the look was more strongly associated with the owner's scolding than with whether the dog had actually disobeyed (Horowitz, 2009). A display that owners confidently read as an admission of wrongdoing turned out to be better explained as a response to the owner's behavior in the moment (what the guilty look actually responds to).

8.5 The Handler Is Part of the Situation

Dogs' facial activity changes with human attention (Kaminski et al., 2017), and the threat test showed that a person's approach style changes how dogs respond (Pedretti et al., 2023). In an adapted still-face paradigm, dogs showed a rich repertoire of visual and facial behaviors during active interaction, more toward their owner than toward a familiar trainer, and these behaviors decreased when the person became unresponsive, whether the person stayed facing the dog or turned away (Canori, Pedretti, Travain, Annoni, Sabbadini & Valsecchi, 2026). Body language observed during an interaction is therefore partly a response to the observer and to the relationship. What a dog shows when a person leans over or stares at it is not a neutral sample, and changing the person's behavior immediately changes part of the social context the dog is responding to (how human states can reach dogs).

9. Practical Implications

9.1 Treat Signals as Information, Not Verdicts

The most defensible use of canine body language is as a running account of how a situation is developing for the dog. An increase in looking away, body lowering, freezing or nose licking during an interaction is a reason to reduce pressure — more distance, less direct approach, a pause — regardless of whether the behavior is ultimately best described as a stress indicator or an appeasement signal.

That recommendation does not depend on resolving the dispute about function, only on the finding that several of these behaviors occur more often in tense situations.

9.2 Respond to Avoidance Early

Head turning and looking away were frequently observed in several studies of tense dog–dog and dog–human encounters. Responding to them by giving the dog space is low-cost and, in a dog that is not inclined to escalate, removes the pressure that might otherwise push it toward stronger responses (what happens when a dog is torn between approaching and avoiding).

9.3 Do Not Suppress Warning Behavior

Punishing a dog for showing avoidance or warning behavior may suppress the visible behavior without resolving the motivation or emotional state that produced it, potentially leaving observers with fewer cues before escalation. The specific sequence from a punished growl to a bite without warning has not been studied directly, but the risk runs in the wrong direction for anyone living or working with the dog (what suppressing warning signals costs). Aversive responses to such behaviors are particularly risky in fearful dogs, where they can intensify the state that drives the problem.

9.4 Teach Observation Before Interpretation

When teaching owners or children, describing what to look for — the head turns, the body lowers, the dog stops moving toward you — is more robust than teaching what each behavior means.

9.5 Know When Body Language Is Not Enough

A dog whose body language changes suddenly or persistently, without an obvious change in its circumstances, may be responding to pain or illness rather than to anything in the social situation. Body language describes a state; it does not diagnose its cause, and a veterinary examination belongs at the start of any unexplained change (how chronic pain changes behavior).

10. Summary at a Glance

Tail wagging is lateralized — Wagging was biased to the right toward the owner and to the left toward an unfamiliar dominant dog in 30 mixed-breed dogs (Quaranta et al., 2007).

Other dogs respond to the asymmetry, but not consistently — Dogs watching left-biased wagging showed higher heart rates and more anxious behavior than dogs watching right-biased wagging (Siniscalchi et al., 2013), while free-ranging dogs approaching a robotic replica stopped more often at right-biased wagging (Artelle et al., 2011).

Faces differ between anticipation and frustration — Ears adductor was more common when a reward was expected; blinking, lips parting, jaw dropping, nose licking and ear flattening when it was withheld, replicated in a new sample (Bremhorst et al., 2019, 2022).

Faces respond to human attention — Dogs produced more facial movements when a person was facing them, while food had no effect (Kaminski et al., 2017).

The inner brow raise is anatomically real, its communicative role unsettled — The muscle was found in five of six dogs dissected but not as a distinct muscle in wolves (Kaminski et al., 2019) and occurs in other canids (Cunningham et al., 2024; Sexton et al., 2024; de Souza-Junior et al., 2026); faster rehoming was associated with the movement (Waller et al., 2013), but it was tied to eye movements rather than to a human audience (Bremhorst et al., 2021).

"Calming signals" are real behaviors with an unconfirmed function — They were frequent in dog–dog encounters and often followed by de-escalation (Mariti et al., 2017), but in a direct test most did not increase under threat (Pedretti et al., 2023).

Blinking elicits blinking — Dogs blinked more while watching videos of dogs blinking than of dogs licking their noses (Canori et al., 2025).

Lip licking has more than one context — It increased in tense situations and was also a common part of greeting people (Firnkes et al., 2017).

Play bows restart play — In 414 bows by adult dogs, bows usually followed a pause and were almost always given within the partner's view (Byosiere et al., 2016).

People miss subtle signs — Owners rarely named head turning, yawning or nose licking as stress signs (Mariti et al., 2012), and young children often read aggressive dog faces as happy (Meints & Racca, 2026).

11. Research Gaps and Critical Appraisal

Function is rarely tested. Most studies show that a behavior occurs more often in one context than another. A direct test of key predictions did not find most putative appeasement behaviors to be specific to the threatening condition (Pedretti et al., 2023), and effects on receivers have so far been tested only for single behaviors such as blinking (Canori et al., 2025).

Observational designs dominate. The dog–dog calming-signal study could not separate a behavior causing de-escalation from a behavior accompanying it (Mariti et al., 2017), and the play bow analysis relied on archived video of a small number of dyads (Byosiere et al., 2016).

Samples are small and narrow. Single breeds, a few dozen dogs and heavily skewed contributions from individual dogs or dyads limit how far the findings generalize.

Independent replication is scarce. The facial-coding associations were reproduced by the same group; most other findings have not been repeated by unconnected researchers. Where a different group tested responses to asymmetric wagging, the direction of the effect was reversed (Artelle et al., 2011).

Real-time applicability is untested. The tail asymmetries and many facial actions were measured with instruments or frame-by-frame coding. Whether people can detect them reliably in live interactions has not been established.

Morphology is barely studied. Tail length changed how other dogs responded to tail motion in one experiment (Leaver & Reimchen, 2008); breed differences in ear shape, coat and facial structure plausibly change what dogs can signal and what observers can see, and none of the studies here measured those effects.

Positive states lack validated indicators. No single indicator has been validated as a reliable, context-independent marker of positive emotion (Csoltova & Mehinagic, 2020); in one controlled study several physiological and behavioral parameters differentiated arousal only within negative scenarios, while qualitative behavioral assessment tracked both valence and arousal (Flint et al., 2024).

Human interpretation studies use simplified stimuli. Photographs and questionnaires measure recognition and awareness, not observation during real encounters (Mariti et al., 2012; Meints & Racca, 2026).

12. Conclusion

Canine body language is neither a code with fixed meanings nor a matter of guesswork. The research shows that dogs' tails, faces and bodies change in consistent ways across situations: wagging is lateralized and other dogs respond to the asymmetry, though not in the same direction in every study (Quaranta et al., 2007; Siniscalchi et al., 2013; Artelle et al., 2011), facial actions differ between anticipation and frustration and increase when a person is watching (Bremhorst et al., 2019, 2022; Kaminski et al., 2017), the inner brow raise rests on a muscle that distinguishes dogs from gray wolves but not from some other canids and is associated with faster rehoming of shelter dogs (Kaminski et al., 2019; Cunningham et al., 2024; Waller et al., 2013), and play bows are given where a partner can see them, usually to restart play (Byosiere et al., 2016). The behaviors popularized as calming signals are frequent, several were associated with a non-reactive behavioral profile, and blinking elicits blinking in watching dogs, but an appeasement function has not been confirmed (Mariti et al., 2017; Firnkes et al., 2017; Pedretti et al., 2023; Canori et al., 2025). At the same time, people notice the conspicuous signs far more readily than the subtle ones, and young children frequently misread aggressive dog facial expressions (Mariti et al., 2012; Meints & Racca, 2026). The practical conclusion follows without resolving every dispute about function: read sequences rather than single behaviors, compare the dog with its own baseline, respond to avoidance by reducing pressure, keep warning behavior intact rather than suppressing it, and treat any list of meanings as a set of hypotheses to be checked against the dog in front of you.

Key Insights (Takeaways)

• Body language research measures behaviors, not meanings. Most findings show that a movement becomes more frequent in some situations; very few show what it communicates or what the dog intends.

• Tail wagging is lateralized — right-biased toward an owner, left-biased toward an unfamiliar dominant dog — and other dogs respond to it, although a video study and a field study with a robotic replica found opposite directions of response (Quaranta et al., 2007; Siniscalchi et al., 2013; Artelle et al., 2011). The asymmetry is hard to see live and does not provide a rule for reading a wagging dog.

• Facial actions differ between positive anticipation and frustration in a replicated experimental design (Bremhorst et al., 2019, 2022), increase when a person is attentive (Kaminski et al., 2017), and in the case of the inner brow raise are associated with how quickly shelter dogs are rehomed (Waller et al., 2013); no single facial action classified the underlying state reliably on its own (Bremhorst et al., 2022).

• "Calming signals" such as head turning, nose licking and looking away are common, and several were associated with a non-reactive behavioral profile, but a direct test of key predictions did not find them to increase under threat as an appeasement function would predict (Mariti et al., 2017; Firnkes et al., 2017; Pedretti et al., 2023), and receiver effects have been shown only for single behaviors such as blinking (Canori et al., 2025). Treat them as information, not as messages with a fixed content.

• In adult dogs, play bows mainly restart play after a pause and are almost always given within the partner's view (Byosiere et al., 2016); a bow is not a guarantee that everything around it is play.

• Owners rarely recognize subtle stress behaviors (Mariti et al., 2012), and young children often read aggressive dog faces as happy (Meints & Racca, 2026). Teaching what to look for is more consistent with the evidence than teaching fixed meanings for individual behaviors, and warning behavior should never be punished away.

References

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

Beerda, B., Schilder, M. B. H., van Hooff, J. A. R. A. M., de Vries, H. W., & Mol, J. A. (1998). Behavioural, saliva cortisol and heart rate responses to different types of stimuli in dogs. Applied Animal Behaviour Science, 58(3–4), 365–381. https://doi.org/10.1016/S0168-1591(97)00145-7

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

Bremhorst, A., Mills, D. S., Stolzlechner, L., Würbel, H., & Riemer, S. (2021). 'Puppy dog eyes' are associated with eye movements, not communication. Frontiers in Psychology, 12, 568935. https://doi.org/10.3389/fpsyg.2021.568935

Bremhorst, A., Mills, D. S., Würbel, H., & Riemer, S. (2022). Evaluating the accuracy of facial expressions as emotion indicators across contexts in dogs. Animal Cognition, 25(1), 121–136. https://doi.org/10.1007/s10071-021-01532-1

Bremhorst, A., Sutter, N. A., Würbel, H., Mills, D. S., & Riemer, S. (2019). Differences in facial expressions during positive anticipation and frustration in dogs awaiting a reward. Scientific Reports, 9, 19312. https://doi.org/10.1038/s41598-019-55714-6

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., Pedretti, G., Travain, T., Annoni, C., Sabbadini, L., & Valsecchi, P. (2026). Can you read my poker-face? Adapting the still-face paradigm to explore dog's interspecific communication. Animal Cognition, 29, 34. https://doi.org/10.1007/s10071-026-02059-z

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

Csoltova, E., & Mehinagic, E. (2020). Where do we stand in the domestic dog (Canis familiaris) positive-emotion assessment: A state-of-the-art review and future directions. Frontiers in Psychology, 11, 2131. https://doi.org/10.3389/fpsyg.2020.02131

Cunningham, P., Shankar, M., vonHoldt, B., Brzeski, K. E., & Kienle, S. S. (2024). Coyotes can do 'puppy dog eyes' too: Comparing interspecific variation in Canis facial expression muscles. Royal Society Open Science, 11(10), 241046. https://doi.org/10.1098/rsos.241046

de Souza-Junior, P., Souza, E. C., Viotto-Souza, W., Costa, M. E. R., Crivelaro, A. Z., Kasper, C. B., & Abidu-Figueiredo, M. (2026). Beyond domestication: Occurrence of levator anguli oculi medialis and retractor anguli oculi lateralis in four Neotropical canid species. Anatomia, Histologia, Embryologia, 55(3), e70121. https://doi.org/10.1111/ahe.70121

Firnkes, A., Bartels, A., Bidoli, E., & Erhard, M. (2017). Appeasement signals used by dogs during dog–human communication. Journal of Veterinary Behavior, 19, 35–44. https://doi.org/10.1016/j.jveb.2016.12.012

Flint, H. E., Weller, J. E., Parry-Howells, N., Ellerby, Z. W., McKay, S. L., & King, T. (2024). Evaluation of indicators of acute emotional states in dogs. Scientific Reports, 14(1), 6406. https://doi.org/10.1038/s41598-024-56859-9

Horowitz, A. (2009). Disambiguating the "guilty look": Salient prompts to a familiar dog behaviour. Behavioural Processes, 81(3), 447–452. https://doi.org/10.1016/j.beproc.2009.03.014

Kaminski, J., Hynds, J., Morris, P., & Waller, B. M. (2017). Human attention affects facial expressions in domestic dogs. Scientific Reports, 7, 12914. https://doi.org/10.1038/s41598-017-12781-x

Kaminski, J., Waller, B. M., Diogo, R., Hartstone-Rose, A., & Burrows, A. M. (2019). Evolution of facial muscle anatomy in dogs. Proceedings of the National Academy of Sciences, 116(29), 14677–14681. https://doi.org/10.1073/pnas.1820653116

Leaver, S. D. A., & Reimchen, T. E. (2008). Behavioural responses of Canis familiaris to different tail lengths of a remotely-controlled life-size dog replica. Behaviour, 145(3), 377–390. https://doi.org/10.1163/156853908783402894

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 (Canis familiaris): 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

Mariti, C., Gazzano, A., Moore, J. L., Baragli, P., Chelli, L., & Sighieri, C. (2012). Perception of dogs' stress by their owners. Journal of Veterinary Behavior, 7(4), 213–219. https://doi.org/10.1016/j.jveb.2011.09.004

Meints, K., & Racca, A. (2026). Look, he's smiling! Children's misinterpretation of dogs' facial expressions. Anthrozoös, 39(2), 315–342. https://doi.org/10.1080/08927936.2025.2551434

Overall, K. L. (2017). Appeasement, calming signals, and information capture: How do our subjects tell us what matters to them? Journal of Veterinary Behavior, 19, v–viii. https://doi.org/10.1016/j.jveb.2017.04.001

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

Pedretti, G., Canori, C., Costantini, E., Palme, R., Valsecchi, P., & Marshall-Pescini, S. (2024). Intra and interspecific audience effect on domestic dogs' behavioural displays and facial expressions. Scientific Reports, 14, 9546. https://doi.org/10.1038/s41598-024-58757-6

Pedretti, G., Canori, C., Marshall-Pescini, S., Palme, R., Pelosi, A., & Valsecchi, P. (2022). Audience effect on domestic dogs' behavioural displays and facial expressions. Scientific Reports, 12, 9747. https://doi.org/10.1038/s41598-022-13566-7

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

Rugaas, T. (2006). On talking terms with dogs: Calming signals. Legacy By Mail.

Sexton, C. L., Diogo, R., Subiaul, F., & Bradley, B. J. (2024). Raising an eye at facial muscle morphology in canids. Biology, 13(5), 290. https://doi.org/10.3390/biology13050290

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

van der Borg, J. A. M., Schilder, M. B. H., Vinke, C. M., & de Vries, H. (2015). Dominance in domestic dogs: A quantitative analysis of its behavioural measures. PLoS ONE, 10(8), e0133978. https://doi.org/10.1371/journal.pone.0133978

Waller, B. M., Peirce, K., Caeiro, C. C., Scheider, L., Burrows, A. M., McCune, S., & Kaminski, J. (2013). Paedomorphic facial expressions give dogs a selective advantage. PLoS ONE, 8(12), e82686. https://doi.org/10.1371/journal.pone.0082686