How Dogs Understand Us: Words, Gestures, Gaze and Tone of Voice

In brief

Four things determine good communication with your dog. First: dogs do not understand human language the way humans do. But they can distinguish learned words and, for some words, apparently form mental representations of what is meant; at the same time, gestures, tone of voice, context and learning history play an important role. Second: what a dog gives more weight to – word, gesture, gaze or tone of voice – depends on the task, learning history and situation. Often several sources of information are processed at the same time. Third: many supposed “disobedience” problems are misunderstandings – inconsistent cues, inconsistency, humanizing or unintentional reinforcement. Fourth: communicating clearly, consistently and with awareness of context works best. In an object-choice study with 20 dogs, the combination of a pointing gesture and a look toward the target was particularly effective.

Dogs do not understand human language the way humans do. They learn numerous cues and connections through experience – they watch our body language, respond to the sound of our voice and orient themselves by recurring patterns. For some familiar words, there is also evidence that they can activate a mental image of what is meant. How exactly this works – and how we can do better – is what this article is about.

👉 The basics: how to read your dog’s body language correctly

Dogs communicate mainly through body language, voice and emotional signals – words are only part of it. Research shows that they process our cues in more complex ways than was long assumed.

The relationship between humans and dogs is not based on language in the classic sense, but on sending and interpreting signals. But how exactly does that work?

Research has made clear progress on this in recent years. One notable study is by Völter and colleagues (2025): using mobile eye tracking, they analyzed how dogs perceive human cues – pointing gestures and gaze direction. The results offer clues about how cues can be made clearer in everyday life and in training.

1. The basics of human-dog communication

Communication with your dog runs on several levels. Body language, voice and timing are especially important.

Body language – dogs read us like a book

Dogs are masters of nonverbal communication. They orient strongly toward our posture, direction of movement and facial expressions. Even small changes in body tension or movement can carry a clear message for them. Relevant factors include the direction our shoulders or feet are pointing, the tension in our body, where we are looking and the dynamics of our movement. Understanding your own body language therefore starts with becoming aware of your own signals.

In practice: When a visitor wants to greet the young dog Cara, he leans over her from the front and from above – well meant, but Cara backs away. Leaning over a dog head-on can be unpleasant for some dogs. As soon as the same person crouches down and turns sideways, Cara relaxes and comes closer on her own. For the dog, it is not the intention that counts but the posture that reaches her.

Voice and tone – additional information

Dogs don’t understand language the way we do, but they are very sensitive to the emotional sound of the voice – to pitch, volume, rhythm and coloring. A “Good!” spoken in a friendly way and learned as something positive can have a confirming effect; a sharp tone conveys different, additional information. Dogs can process word content and intonation separately. Which information – word, tone, gesture or context – carries more weight in a specific situation depends on the task and learning history.

Timing – the right moment makes the difference

For a dog to learn, feedback has to come at the right moment. Especially when building a new behavior, the feedback should be as immediate as possible. A well-established marker signal can precisely mark the decisive moment and bridge a short delay until the actual reward. Precise timing makes it easier for the dog to connect behavior and consequence. With established cues, the time window can become more generous because the behavior has already been learned – but good timing remains a key factor.

2. Common misunderstandings

Many everyday problems arise not because dogs “don’t want to listen,” but because they interpret our signals differently from how we meant them.

Inconsistent cues: If different words or gestures are used for the same behavior – “Here,” “Come,” “Come here” – it is hard for the dog to recognize what is meant. Clear, consistent cues make learning enormously easier.

Unclear conditions: Changing consequences without a recognizable distinguishing feature can make learning harder. The classic example: the couch. Sometimes the dog is allowed up, sometimes not – without a signal that marks the difference, the dog cannot tell which rule applies right now. With a clear distinguishing feature, on the other hand, dogs learn rules like this very well.

Humanizing: Questions like “Why did you do that?” or moral reproaches cannot be understood by dogs. They respond to current signals and consequences, not to evaluations after the fact. As for the much-cited “guilty look”: in a study of 14 dogs, it was more closely linked to the owner scolding than to whether the dog had actually carried out the forbidden action (Horowitz, 2009). So the look is not reliable evidence that a dog is evaluating its earlier action – but it does not follow that dogs fundamentally have no guilty conscience.

Unintentional reinforcement: A dog barks at the front door, and the person scolds or shouts. It is meant as a reprimand. If a person responds to barking every time, that response can maintain the behavior in some dogs – provided the attention actually works as a reinforcer for that dog. Whether scolding is reinforcing, irrelevant or unpleasant is decided not by our intention but by its effect on the behavior.

In practice: Wanda barks every time the doorbell rings, and her person then shouts “Quiet!” down the hallway. The barking continues. Only when he calmly turns the doorbell into the cue for a hand target and a reward instead does the barking lose its purpose.

3. What research shows: the study by Völter and colleagues

The study “Using mobile eye tracking to study dogs' understanding of human referential communication” by Christoph J. Völter and colleagues (2025) offers revealing insight. What is new is the method: the dogs wore a lightweight mobile eye-tracking system that recorded where they were looking during direct human-dog interaction in a controlled object-choice task.

Five conditions were tested, each indicating the location of hidden food: pointing, pointing combined with gaze, gaze only, a directional fake-throw movement, and a control with no cue at all. The question behind it: do dogs respond differently to communicative, referential signals than to a merely directional movement?

The result: the combination of a pointing gesture and a look toward the target was particularly effective in this task. The comparison with the throwing gesture was especially revealing – only with genuine pointing did the dogs direct their gaze specifically from the hand to the indicated bowl, not in the same way with the mere throwing movement. This suggests that dogs respond specifically to communicative cues and do not simply follow any movement.

The authors themselves are careful here: whether dogs truly understand such gestures as “referential communication” in the human sense, or respond to them in a learned or reflexive way, remains scientifically open. How far this goes: 20 pet dogs were analyzed, and an object-choice task was tested. So the finding is: in this referential task, pointing combined with a look toward the target was particularly effective. Recall, sit, voice combined with hand signals, or training in general were not studied – it does not follow that multiple cues would be better in every training situation.

4. Building blocks of successful communication

A few principles can be derived from learning psychology and recent studies.

Clarity: Cues should be unambiguous and consistent, and body language and words should not contradict each other. In the Völter study, the combination of pointing and looking toward the target was particularly effective in the referential task examined.

Reliability: Dogs need to be able to rely on cues and rules staying stable. Consistent cues and recognizable rules make it easier for your dog to predict which behavior will lead to success in a situation – and so they also make working together easier.

Positive reinforcement: Desired behavior is consistently confirmed – with food, play or social interaction, depending on what motivates this particular dog. This is exactly where modern, force-free training comes in; on why aversive methods are the wrong path, see Aversive Training Methods in Dogs: Neurological Effects, Stress Responses and Long-Term Welfare Risks.

Awareness of context: Communication always takes place in a context. Studies of adult dogs show that they can use emotional information from their caregiver to assess ambiguous situations (Merola et al., 2012).

In practice: When calling him back, Tomte’s person often just shouts his name across the meadow in an irritated voice – and Tomte dawdles. As soon as the person turns toward the dog, crouches down, calls clearly and opens his arms invitingly, Tomte comes reliably. The same intention, but sent much more clearly.

5. How much do dogs actually understand of words?

The blanket statement “dogs don’t understand words” would be too crude. Research now paints a more nuanced picture.

5.1 Word content and tone of voice are processed separately

An imaging study showed back in 2016 that dogs process lexical and intonational information separately: meaningful words produced different neural responses from meaningless ones – regardless of tone. Reward regions responded especially when praising word content and praising intonation matched (Andics et al., 2016).

So for a dog, a word is not just a sound that the tone of voice happens to be attached to.

5.2 Words can refer to objects

An EEG study from 2024 goes one step further: when a dog heard a familiar object word and was then shown the wrong object, a response occurred that resembles the human N400 response to semantic violations. The authors interpret this as an indication that familiar object words can activate a mental image of the object named (Boros et al., 2024).

A more recent imaging study points in the same direction: for action and instruction words, neural representations were found that go beyond merely recognizing the sound form and are compatible with a representation of meaning (Boros et al., 2025). This means the finding concerns not only object words but also the cues actually used in training.

5.3 Large vocabularies are still rare

This needs to be distinguished from the ability to deliberately learn many new object names and retrieve them on request. In one study, 34 previously inexperienced family dogs and six dogs already known for their unusual word knowledge were trained for three months: only one of the 34 ordinary dogs successfully learned the object names, while all six exceptional dogs did (Fugazza et al., 2021).

For context: Exceptionally large object vocabularies and fast learning of new names are rare. However, the more recent EEG research suggests that a referential understanding of individual familiar object words may also occur in less exceptional dogs – retrieval tasks may underestimate it.

5.4 What this means for everyday life

When building a cue, a clear, consistent signal helps first. As soon as the behavior is understood, you should deliberately train in different body positions, situations and environments – otherwise your dog may end up responding to unintended accompanying signals instead of the word.

6. What dogs read in us – and what we overlook

Section 1 names the channels. In everyday life, though, something else usually makes the difference: we send signals all the time without noticing.

6.1 The unconscious signals

The leash. A shortened or tight leash immediately changes your dog’s freedom of movement and the physical feedback he gets. If you close your hand more tightly when meeting another dog, you have already changed something – regardless of what you say.

The routine beforehand. Keys, shoes, jacket: your dog learns the sequence, not just the individual cue. That’s why he is already standing at the door before you have said anything.

That is why it pays to keep an eye on such accompanying stimuli as well as your deliberate cues – especially when a cue doesn’t work as usual in a particular situation.

6.2 Subtle changes are easily overlooked

The same applies the other way around. Looking away, licking the nose, briefly freezing or yawning can – depending on the context – fit a changed emotional state, and these are exactly the signs that can easily be overlooked.

Important context: None of these signals proves stress on its own. What matters is the context, how often they occur, and how far they deviate from your dog’s normal behavior – yawning can be tiredness, licking can have to do with food.

If you see these signs early, you can adjust a situation sooner. That is the real gain: communication runs in both directions, and half of it is listening.

7. When a cue doesn’t work

A common conclusion is: “He doesn’t listen.” Four common training questions help with troubleshooting. Other causes are also possible – pain, hearing or vision problems, tiredness, fear or a competing motivation.

7.1 The cue was never built properly

A word that was used ambiguously from the start stays ambiguous as long as it is used that way. “No” for barking, pulling, jumping up and picking up things that aren’t his is not a cue but a noise.

7.2 It only works in one place

Dogs do not automatically transfer what they have learned from one situation to another. A sit that works in the living room often doesn’t work outside yet – that is normal, not a setback.

7.3 The situation is too much right now

Under high arousal, cues are harder to retrieve. The starting point then is not the word but the distance to the trigger.

7.4 It no longer pays off

A cue that has had no consequences for months can lose its value. This applies especially to recall when it only announces the end of freedom.

And what rarely helps: getting louder or repeating the cue. If a cue is regularly repeated several times before a response is required or reinforced, a response may come to occur only after several repetitions. If a cue doesn’t get through, you are better off taking a step back than repeating it louder.

8. From everyday training practice

Day-to-day work shows again and again: dogs do not respond randomly but logically to what they perceive. Many problems arise not from a lack of obedience but from unclear communication. As soon as owners learn to use their cues and their own body language deliberately and to better understand how their dog thinks and perceives, the whole relationship often changes – and the behavior along with it.

9. Conclusion

Dogs do not understand us in the same way other people do, but through a mix of word, gesture, tone of voice and context. With some training problems, it is therefore worth looking not only at the dog’s behavior but also at your own signals, the context and the learning history so far. That requires attention and a willingness to learn on the human side – because our dogs do perceive complex signals, provided we send them clearly, consistently and with awareness of context. If you don’t just “train” your dog but communicate with him, you lay the foundation for trust, cooperation and a stable human-dog team.

References

• Völter, C. J., Gerwisch, K., Berg, P., Virányi, Z. & Huber, L. (2025). Using mobile eye tracking to study dogs' understanding of human referential communication. Proceedings of the Royal Society B, 292(2040), 20242765. https://doi.org/10.1098/rspb.2024.2765
• Andics, A., et al. (2016). Neural mechanisms for lexical processing in dogs. Science, 353(6303), 1030–1032. https://doi.org/10.1126/science.aaf3777
• Boros, M., et al. (2024). Neural evidence for referential understanding of object words in dogs. Current Biology, 34(8), 1750–1754.e4. https://doi.org/10.1016/j.cub.2024.02.029
• Boros, M., Rácz, K. & Andics, A. (2025). Action instruction word processing in the dog brain entails both auditory form identification and meaning representation. NeuroImage, 318, 121411. https://doi.org/10.1016/j.neuroimage.2025.121411
• Fugazza, C., et al. (2021). Word learning dogs provide an animal model for studying exceptional performance. Scientific Reports, 11, 14070. https://doi.org/10.1038/s41598-021-93581-2
• 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
• Merola, I., Prato-Previde, E. & Marshall-Pescini, S. (2012). Dogs' social referencing towards owners and strangers. PLOS ONE, 7(10), e47653. https://doi.org/10.1371/journal.pone.0047653

Key takeaways on communication

Dogs don’t understand language the way humans do – but words are more than noise to them. Word content and tone of voice are processed separately (Andics et al., 2016), and familiar object words can apparently activate an image of what is meant (Boros et al., 2024). For action and instruction words, too, representations were found that go beyond recognizing the sound form (Boros et al., 2025).

Large object vocabularies remain rare. In one study, only one of 34 ordinary family dogs learned new names, while all six dogs already known for their word knowledge did (Fugazza et al., 2021) – an understanding of individual familiar words may still be more widespread.

There is no fixed ranking between word, gesture, gaze and tone of voice. What a dog gives more weight to depends on the task, learning history and situation – often several sources of information are processed at the same time.

Many supposed obedience problems are misunderstandings – inconsistent cues, changing consequences or unintentional reinforcement.

Unintended accompanying stimuli have an effect too: a tight leash changes freedom of movement and physical feedback, and your dog learns the sequence of your routines, not just the individual cue.

Subtle changes are easily overlooked – looking away, licking the nose, briefly freezing, yawning. None of these proves stress on its own; what matters is context and frequency.

Four training questions help with troubleshooting: Was it built properly? Was it practiced in several places? Is the situation too hard right now? Does it still pay off? Pain, sensory problems, tiredness or competing motivation are also possible causes.

If a cue is regularly repeated several times before a response is required, the response may come to occur only after several repetitions. Better to take a step back than to get louder.

What has been learned does not automatically transfer from one place to another. That is normal, not a setback.

Communication runs in both directions – and half of it is listening.

Frequently asked questions about communicating with dogs

How your dog really understands you – and how you can improve your cues