Behavior Does Not Equal Emotion: Limits of Inferring Internal States from Observable Actions

2. Why Behavior Is Not a Direct Readout of Emotion

Behavior is shaped by multiple factors, only one of which is the animal’s current emotional state. Other major influences include:

• Learning history – A dog may have been reinforced for hiding signs of fear (e.g., punished for growling) and therefore learns to suppress emotional expressions while still feeling afraid.

• Context – The same behavior (e.g., lying down) can indicate relaxation in one environment and learned helplessness or cautious monitoring in another.

• Physiological state – Pain, fatigue, hunger, or illness can alter behavior independently of emotion.

• Conflicting motivations – A dog may simultaneously want to approach (curiosity, social motivation) and avoid (fear), producing mixed or ambiguous behaviors.

• Individual differences – Breed, temperament, age, and prior experience all affect how emotions are expressed behaviorally.

The fundamental problem is underdetermination: the same observable behavior can arise from different internal states, and the same internal state can produce different behaviors depending on context and learning history.

A conceptual model:

Behavior ≠ Emotion. The relationship between the two is moderated by at least three interacting factors:

• Learning history (e.g., reinforcement/punishment of emotional expression)

• Context (environmental and social setting)

• Physiological state (pain, arousal, hunger, illness)

Thus, inferring emotion from behavior requires information beyond the behavior itself.

For a deeper look at how learning history shapes behavior independently of emotion, see learned behavior vs. emotional response in dogs.

3. The Dissociation Between Behavior and Emotion – Key Examples

3.1 Tail Wagging: Not Always Happiness

Tail wagging is perhaps the most frequently misinterpreted canine behavior. While a loose, sweeping, whole‑body wag often indicates positive arousal, tail wagging also occurs in contexts of uncertainty, anxiety, or impending aggression.

Research has shown that:

• The asymmetry of tail wagging (biased to the left or right) may correlate with different emotional valences (left‑biased wagging associated with approach/positive, right‑biased with withdrawal/negative) (Siniscalchi et al., 2013), but this is not easily observable in real‑time by owners.

• High, stiff, rapid wagging (often described as “flagging”) is more indicative of high arousal, which could be excitement, fear, or aggression.

• Dogs may wag their tails immediately before biting – not as a sign of friendliness, but as a component of high arousal that includes both approach and threat.

Conclusion: Tail position, speed, and stiffness matter, but even with these cues, tail wagging alone is insufficient to determine emotion.

3.2 Stillness: Relaxation or Freezing?

A dog that lies still may be relaxed – or may be freezing. Freezing is a fear response characterized by immobility, shallow breathing, and hypervigilance. It is an active coping strategy in which the dog attempts to become invisible to a perceived threat. Freezing is distinct from relaxation (low arousal, positive or neutral valence) and from learned helplessness (a passive state resulting from repeated uncontrollable aversive events).

Distinguishing relaxation from freezing requires attention to subtle cues:

• Muscle tone – Relaxed dogs have loose, soft muscles; freezing involves increased muscle tension.

• Breathing – Relaxed breathing is slow and regular; freezing often involves shallow, rapid breaths or breath‑holding.

• Eyes – Relaxed eyes are soft, with normal blinking; freezing often involves wide eyes, dilated pupils, and reduced blinking (“whale eye” – showing the sclera).

• Responsiveness – A relaxed dog may respond to its name or a treat; a freezing dog may be unresponsive or slow to respond.

A dog that has been trained to “down‑stay” may remain still because of operant conditioning, not because it is calm. The stillness is behavior – the emotion could range from contentment to anxiety.

3.3 Approach: Friendly or Fear‑Related?

Approaching a person or another dog is often interpreted as sociability. However, approach can occur alongside fear, creating approach‑avoidance conflict. The dog may approach slowly, with a tucked tail, ears back, and a tense body – still “approaching,” but motivated by social pressure, curiosity mixed with caution, or appeasement.

Appeasement refers to behaviors (e.g., lip licking, turning away, lowering the body) that function to reduce perceived threat from another individual. Appeasement can coexist with approach, making the dog appear “friendly” when it is actually anxious.

A 2024 study on dogs’ responses to human stress odor found that dogs exposed to stress odor were less likely to approach an ambiguous location – but dogs that did approach showed no difference in behavior, despite having different internal states (as measured by cognitive bias). The behavior looked the same; the emotion differed.

3.4 Vocalization: Barking, Whining, Growling

Vocalizations are often assumed to map directly onto emotions: growling = aggression, whining = anxiety, barking = excitement. In reality, vocalizations are context‑dependent and can be ambiguous (Pongrácz et al., 2006).

• Growling can occur in play, in resource guarding, in fear, or in pain. Play growls are typically higher‑pitched and occur alongside play bows and loose body language; fear growls are often lower, accompanied by a tense body and lip licking.

• Whining can indicate anxiety, frustration, anticipation, excitement, or even pain.

• Barking varies in pitch, duration, and rhythm, but even experienced owners often misclassify barks by emotion.

For more on the neurological basis of fear and aggression, see reactivity in dogs – a neurological perspective.

4. Masking and Suppression – When Behavior Hides Emotion

One of the most important concepts in understanding the behavior–emotion gap is masking (also called suppression): the dog learns to inhibit outward signs of an emotional state while still experiencing that state internally. Masking is distinct from genuine emotional regulation (where the internal state changes) and from inhibitory learning (where a new memory competes with the original, as in extinction).

4.1 How Masking Develops

Masking typically results from learning history, particularly punishment of emotional expressions:

• A dog that growls when afraid may be punished (e.g., scolded, hit, or corrected with a leash pop). The growling stops, but the fear remains – or intensifies, because the punishment confirms that the threat was real.

• The dog learns that warning signals are dangerous, so it stops giving them. The dog may go directly from no visible warning to a bite (so‑called “bite out of nowhere”).

• A dog that learns to perform a “calm” behavior (e.g., lying down on a mat) to receive a treat may appear relaxed while internally aroused. The behavior is reinforced; the emotion is not addressed.

4.2 Masking vs. Emotional Regulation vs. Learned Helplessness

It is important to distinguish three related but distinct phenomena:

• Masking (suppression) – The outward expression of emotion is inhibited, but the internal state remains unchanged or worsens. The dog has learned that displaying emotion leads to punishment.

• Emotional regulation – The internal state itself changes (e.g., fear decreases through counter‑conditioning). The dog is genuinely calmer.

• Learned helplessness – A passive state resulting from repeated exposure to uncontrollable aversive events. The dog stops trying to avoid or cope; immobility is not relaxation but a breakdown of active coping.

Masked dogs are often described as “shut down” – shut down refers to a state of behavioral inhibition and reduced responsiveness often seen after chronic stress or punishment, where the dog has stopped offering behaviors and appears “well‑behaved” but is not calm.

4.3 Why Masking Is Dangerous

Masking is dangerous for several reasons:

• False assessment – Owners and trainers may believe the dog is calm and push it into situations it cannot handle, leading to sudden escalation.

• Loss of warning signals – The dog no longer growls, snarls, or shows teeth before biting, increasing bite risk.

• Chronic stress – Suppressing emotional expression without resolving the underlying emotion is stressful, leading to elevated cortisol and long‑term welfare impairment (Beerda et al., 1998).

• Treatment failure – Behavior modification that targets only the observable behavior (e.g., training “quiet” without addressing fear) may produce masking, not resolution.

For more on the effects of aversive methods that encourage masking, see aversive training methods – neurological effects in dogs. For chronic stress from unresolved emotional conflict, see neurobiology of chronic stress in dogs – cortisol impact.

5. Arousal and Valence – Two Independent Dimensions of Emotion

Modern affective neuroscience (Mendl et al., 2010) distinguishes between two independent dimensions of emotion:

• Arousal – The intensity of the emotional response (low to high). Arousal reflects physiological activation, including heart rate, cortisol, and sympathetic nervous system activity.

• Valence – The pleasantness or unpleasantness of the emotion (negative to positive).

Crucially, arousal and valence are orthogonal. A dog can be:

• Low arousal, positive valence – Calm, content, relaxed.

• Low arousal, negative valence – Bored, depressed, lethargic, or in a state of learned helplessness.

• High arousal, positive valence – Excited, playful, eager.

• High arousal, negative valence – Fearful, anxious, aggressive, frustrated.

Behavioral observation alone often fails to distinguish between high‑arousal states of opposite valence. A dog that is frantically barking at the door could be excited (positive) or terrified (negative). 

A dog that is jumping up could be greeting joyfully or trying to escape a perceived threat.

Practical implication: To assess emotion, trainers must measure or infer both arousal (using physiological indicators like heart rate, cortisol, or pupil dilation) and valence (using approach/avoidance, cognitive bias, or facial expression). Behavior alone is insufficient.

Recent neuroimaging studies in dogs (e.g., fMRI of reward and threat processing) support this dimensional model, showing distinct neural circuits for arousal and valence that do not always produce unique behavioral outputs.

For a detailed discussion of how arousal affects learning, see arousal regulation in dogs – neurophysiology and training.

6. Common Misinterpretations in Everyday Dog Training and Ownership

6.1 “He Knows He Did Something Wrong” (Guilt)

Perhaps the most widespread misinterpretation is the so‑called “guilty look” – ears back, head down, tucked tail, avoiding eye contact. Owners interpret this as the dog knowing it misbehaved (e.g., chewed a shoe). Research by Horowitz (2009) showed that this posture is actually a response to the owner’s angry or disappointed tone and body language, not an awareness of past misdeeds. Dogs show the “guilty look” even when they have not misbehaved, if the owner believes they have. The behavior is appeasement, not guilt.

6.2 “He’s Calm Now” (After Correction)

A dog that stops pulling on the leash after a sharp correction or stops barking after a scolding is often described as “calm.” In many cases, the dog has not become calm – it has become suppressed. Heart rate, cortisol, and muscle tension may remain elevated. The stillness may reflect fear‑based inhibition or learned helplessness, not relaxation.

6.3 “He Loves Hugs” (Tolerance vs. Preference)

Many dogs tolerate hugs from their owners without attempting to escape. This is often interpreted as enjoyment. However, behavioral studies show that most dogs show signs of stress during hugs (lip licking, ear position changes, turning head away, whale eye), even if they do not actively resist. Tolerance is not the same as preference. Dogs have evolved to seek physical proximity but not necessarily close body restraint.

6.4 “He’s Friendly” (Based on Tail Wagging Alone)

As discussed, tail wagging alone is an unreliable indicator of friendliness. Dogs that bite children or other dogs are often described afterward as “but he was wagging his tail!” The wagging was present – but it indicated high arousal, not positive valence.

For a deeper look at separation‑related behaviors that are often misinterpreted, see separation anxiety in dogs – neurobiology and panic and neurobiology of separation anxiety in dogs.

7. How to Assess Emotional States More Accurately

Given the limits of behavioral observation alone, how can owners and trainers assess a dog’s internal state?

7.1 Multi‑Channel Assessment

Use multiple sources of information:

• Behavioral context – What happened immediately before? What is the dog’s learning history in similar situations?

• Body language (beyond single cues) – Look at the whole body: ear position, eye shape, mouth tension, tail position and movement, body posture, respiratory rate.

• Physiological indicators – Heart rate, panting (non‑thermal), pupil dilation, cortisol levels (in research settings). Portable heart rate monitors and behavioral coding of stress indicators are increasingly accessible.

• Cognitive bias tests – How does the dog respond to ambiguous stimuli? Pessimistic bias suggests negative valence (Mendl et al., 2010; recent studies in shelter and pet dogs).

• Choice and agency – What does the dog do when given freedom to approach or avoid? Approach/avoidance balance is a powerful indicator of valence.

7.2 The “Consent Test” (Preference Assessment)

One practical method for assessing emotional valence is the consent test: the owner stops the interaction (e.g., petting, hugging, playing) and observes whether the dog re‑engages voluntarily. If the dog moves away, turns its head, or disengages, the interaction was likely not positively valenced. If the dog re‑approaches, nudges the hand, or signals for more, the interaction is likely enjoyable.

7.3 Behavioral Indicators of Positive Valence (Beyond Tail Wagging)

Reliable indicators of positive valence include:

• Soft, relaxed body posture

• Loose, wiggly, whole‑body movement

• Play bows

• Relaxed, open mouth with tongue slightly out

• Normal blinking, soft eyes

• Seeking proximity voluntarily

• Tail wagging in a wide, low, sweeping arc

7.4 Indicators of Negative Valence or High Arousal (Mixed Valence)

• Tense, stiff body

• Rapid, shallow breathing

• Dilated pupils

• Lip licking, yawning (in non‑fatigued, non‑hungry contexts) – often displacement behaviors

• Whale eye (showing sclera)

• Ears back or flattened

• Tail tucked, low, or held high and stiff

• Freezing or slow movement

• Displacement behaviors (scratching, sniffing, shaking off without being wet)

For a framework that integrates emotional assessment with training, see learned behavior vs. emotional response in dogs.

8. The Welfare Implications of Misreading Emotion

Misinterpreting behavior as emotion is not merely an academic error. It has direct welfare consequences:

• Pushing a fearful dog into situations – Believing a still dog is calm leads owners to expose the dog to triggers that should be avoided, worsening fear and potentially causing defensive aggression.

• Punishing warning signals – A dog that is punished for growling learns not to growl. The fear remains, and the next step may be a bite without warning.

• Missing pain – A dog that is irritable or avoids touch may be labeled “grumpy” or “dominant” when in fact it is in pain. The underlying medical issue goes untreated.

• Using aversive methods – Owners who believe a dog is “stubborn” or “choosing to misbehave” are more likely to use punishment. If the dog is actually fearful, punishment confirms the fear and worsens the problem.

• Delaying appropriate treatment – Masked emotional states (e.g., a “calm” dog who is actually shut down) may not receive behavioral or pharmacological help because the problem is not visible.

For more on pain as a hidden cause of behavioral change, see chronic pain and aggression in dogs – osteoarthritis and visceral pain in dogs – behavior.

9. Practical Recommendations for Trainers and Owners

Do not assume single‑cue correspondence – A wagging tail, a stillness, or a growl is not a direct readout of a single emotion. Assess the whole body and context.

Look for multiple, concordant cues – Reliability increases when multiple indicators point to the same valence (e.g., soft body + loose wag + re‑engagement).

Use the consent test – Stop the interaction and observe whether the dog chooses to continue. This reveals valence more clearly than passive observation.

Distinguish between suppression and emotional change – If a behavior stops after punishment, ask: has the emotion changed, or has the dog learned to hide it? If in doubt, assume masking.

Measure arousal separately from valence – High arousal (panting, pacing, dilated pupils) indicates intensity but not whether the dog is excited or afraid. Do not mistake high arousal for happiness.

Rule out medical causes first – Before attributing behavior to emotion, rule out pain, illness, sensory deficits, and neurological conditions.

Train with choice and agency – Allowing the dog to control its exposure to triggers (within safety limits) reveals its genuine preferences and emotional state.

When in doubt, consult a professional – Veterinary behaviorists and qualified applied animal behaviorists use standardized assessment tools (e.g., cognitive bias tests, physiological measurements) to infer internal states more accurately than observation alone.

For a deeper look at how attachment and early experiences shape emotional expression, see attachment styles in dogs – secure, avoidant, and ambivalent and sensitive period in puppies – brain and behavior.

10. Summary Table

Key Distinction: Behavior vs. Emotion

• Behavior: Observable action (sit, wag tail, growl, lie still)

• Emotion: Internal state (fear, joy, anxiety, contentment)

• Relationship: Correlated but not deterministic; moderated by learning history, context, physiology

Arousal (Intensity)

• Definition: Degree of physiological activation (low to high)

• Observable cues: Heart rate, breathing rate, pupil dilation, muscle tension, activity level

• Interpretation: Indicates intensity, not valence. High arousal can be positive or negative.

Valence (Pleasantness)

• Definition: Positive or negative quality of emotion

• Observable cues: Approach vs. avoidance, cognitive bias, facial expression, re‑engagement

• Interpretation: Requires multiple cues; cannot be inferred from arousal alone

Masking (Suppression)

• Definition: Learned inhibition of emotional expression while emotion persists

• Observable cues: Absence of warning signals (e.g., no growl before bite); “shut down” posture with muscle tension

• Risk: False assessment of calm; increased bite risk; chronic stress

Common Misinterpretations

• “Guilty look” → Appeasement to owner’s anger, not guilt (Horowitz, 2009)

• “Calm after correction” → Fear‑based inhibition or learned helplessness, not relaxation

• “He likes hugs” → Tolerance, not necessarily preference

• “Tail wagging = friendly” → High arousal, could be positive or negative valence

Key Insights (Takeaways)

• Behavior and emotion are correlated, not identical – The same observable action can arise from different internal states, and the same internal state can produce different behaviors. The relationship is moderated by learning history, context, and physiology.

• Arousal (intensity) and valence (pleasantness) are independent dimensions – High arousal can indicate excitement or fear; low arousal can indicate contentment or depression/learned helplessness.

• Masking (suppression) is common in dogs punished for emotional expression – The dog stops growling or moving but remains afraid; this increases bite risk and chronic stress.

• Tail wagging is not a reliable happiness indicator – Wagging indicates high arousal; valence must be assessed from other cues (body posture, ear position, eye shape, context).

• Stillness is ambiguous – A dog may be relaxed, freezing (fear), performing a trained down‑stay, or in a state of learned helplessness.

• The “guilty look” is appeasement, not guilt – Dogs show this posture in response to owner anger, regardless of whether they misbehaved (Horowitz, 2009).

• Multi‑channel assessment (body language + physiology + context + choice) is essential – No single cue reliably reveals emotion.

• Misreading emotion harms welfare – It leads to pushing fearful dogs into triggers, punishing warning signals, missing pain, delaying treatment, and using aversive methods that worsen the underlying problem.

Conclusion

Behavior is the visible tip of an invisible iceberg. What a dog does – wagging, lying still, growling, approaching – is shaped by multiple factors, only one of which is its internal emotional state. Learning history, context, physiological state, conflicting motivations, and individual differences all influence the expression of emotion. The same behavior can arise from fear, excitement, anxiety, or contentment. The same emotion can produce barking, freezing, or fleeing, depending on the dog’s history and options.

This does not mean behavior is useless for assessing emotion. On the contrary, behavior is our primary window into the canine mind – but it is a window, not a direct feed. Accurate assessment requires multiple, concordant cues, attention to context, knowledge of learning history, and, where possible, physiological indicators or choice tests.

When you see a behavior, pause before attributing an emotion. Ask not only “What is the dog doing?” but also “What is the dog’s arousal level? What is the likely valence? What alternative explanations exist? Could the dog be masking? What would the dog choose if given freedom?”

By moving beyond simplistic behavior‑emotion equations, trainers and owners can avoid dangerous misinterpretations, improve welfare, and design interventions that address the real underlying state – not just its observable expression. A dog that behaves calmly may still need help. A dog that wags its tail may still be afraid. And a dog that has stopped growling may be closer to biting than ever before.

The goal is not to eliminate emotional expression, but to understand it – and to respond to the emotion, not just the behavior.

References

Beerda, B., Schilder, M. B. H., van Hooff, J. A. R. A. M., & de Vries, H. W. (1998). Behavioural, saliva cortisol and heart rate responses to different types of stimuli in dogs. Applied Animal Behaviour Science, *58*(3-4), 365–381.

Horowitz, A. (2009). Disambiguating the “guilty look”: Salient prompts to a familiar dog behaviour. Behavioural Processes, *81*(3), 447–452.

Kujala, M. V. (2017). Canine emotions as seen through human social cognition. Animal Sentience, *2*(14), 1–24.

Mendl, M., Burman, O. H. P., & Paul, E. S. (2010). An integrative and functional framework for the study of animal emotion and mood. Proceedings of the Royal Society B, *277*(1696), 2895–2904.

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.

Schöberl, I., Beetz, A., Solomon, J., Wedl, M., Gee, N., & Kotrschal, K. (2016). Social factors influencing cortisol modulation in dogs during a strange situation procedure. Journal of Veterinary Behavior, *15*, 1–10.

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*(20), 1999–2002.

Wan, M., Bolger, N., & Champagne, F. A. (2019). Human perception of fear in dogs varies according to experience with dogs. PLoS ONE, *14*(10), e0223576.

Recent sources (2020–2025) – recommended for further reading:

Bunford, N., Gábor, A., & Gácsi, M. (2021). Neural reward response and motor inhibition are related to impulsivity in dogs: An event‑related potential study. Scientific Reports, *11*, Article 15802. (Neuroimaging/ERPs)

Parr‑Cortes, Z., Müller, C. T., Talas, L., Mendl, M., Guest, C., & Rooney, N. J. (2024). The odour of an unfamiliar stressed or relaxed person affects dogs’ responses to a cognitive bias test. Scientific Reports, *14*, Article 15843. (Cognitive bias, emotional contagion)

Travain, T., & Valsecchi, P. (2021). The assessment of emotional states in dogs: A review of the current state of the art. Animals, *11*(10), 2857. (Comprehensive review of assessment methods)