What Your Dog's Guilty Look Can and Can't Tell You

Author

Luboslav

Published

October 5, 2026

Reading time

5 min read

Topics

ContextObservation Method

You find a torn cushion. Your dog lowers its head, avoids your gaze and pauses near the doorway. The scene seems to explain itself: the damage supplies a motive for the expression, and the expression supplies a confession. But those two pieces of information were collected at different times. The cushion tells you something happened earlier. Your dog's posture tells you what the dog is doing now.

Before interpreting the look, separate the questions. What caused the damage? What happened when you discovered it? And what might help prevent the next incident? You can address the household problem without treating a facial expression as a reliable account of the past.

▶Play video

The experiment that separated behaviour from blame

Alexandra Horowitz's 2009 study videotaped 14 dogs across trials involving a forbidden treat. The procedure varied whether dogs disobeyed while their owners were away and what owners believed had happened. Behaviours identified as the guilty look did not reliably distinguish dogs that had eaten the treat from those that had not. They were more common when owners scolded, with a stronger effect among obedient dogs. The findings support a response to owner cues rather than a dependable sign of recognising a misdeed. This was a test of the look, not a demonstration that dogs can never experience guilt. [S1]

What if you haven't said anything yet?

Ostojić, Tkalčić and Clayton examined a related question in 2015. They manipulated whether dogs ate forbidden food and whether evidence of the food was present when owners returned. Owners could not infer eating from their dogs' greetings better than chance. The dogs' greeting behaviour also did not differ according to their action or the presence of the food. These results did not support the proposed guilty-look response without a concurrent negative reaction from the owner. They concern this particular experimental greeting, not every possible expression in a home. [S2]

At home, 'I had not spoken' is a useful observation, but it does not describe your whole entrance. You may have stopped abruptly, looked at an object, changed your posture, or approached differently. Consider those details before declaring the reaction a confession. That is a way to examine the scene more carefully, not a claim that you can identify the exact cue after the fact.

Put the discovery on a timeline

Suppose you left at nine, returned at eleven and found a shoe chewed. You do not know whether the chewing occurred immediately after departure or five minutes before your return. You also do not know, from the doorway scene alone, what your dog was doing just before you arrived. A response to your arrival cannot supply those missing times.

Write down what you actually know: the shoe was accessible, it was damaged during the period you were away, and your dog changed posture when you entered. If a camera already recorded the event, reviewing that footage may answer a timing question. Do not recreate a forbidden-food test or display the damaged object to provoke a reaction. Neither is necessary to plan a better arrangement.

This approach also helps in a home with more than one dog. The individual that hangs back may simply be the one whose reaction catches your attention. Keep evidence about the incident separate from reactions to the subsequent confrontation.

Solve the household problem directly

First, check whether the damaged item presents an immediate concern, such as swallowed material or access to something hazardous. If ingestion is possible, contact a veterinarian rather than waiting for a behavioural explanation. Once the immediate situation is safe, clean up calmly and examine how the item became available.

For the next absence, put shoes and other tempting objects out of reach and provide an appropriate resting area and safe activities your dog already uses comfortably. Choose an arrangement suited to this dog; suddenly imposing unfamiliar confinement is not a substitute for understanding the problem. Management can prevent another damaged shoe while you work out whether the incident was isolated or part of a pattern.

If you catch chewing as it begins, calmly redirect to an appropriate available object and acknowledge that choice. If you discover damage later, plan the next opportunity instead of using the discovery for a delayed confrontation. Make the behaviour you want easy to perform, and make your response clear when it happens.

Repeated destruction deserves more information: which objects, at what times, in whose presence, and with what other behaviour? An isolated shoe and repeated distress during absences are different concerns. If the pattern suggests difficulty coping, or changes suddenly, seek veterinary and qualified behavioural guidance rather than assigning a label such as spite or guilt.

A better question at the doorway

Your first question can be 'What needs attention now?' rather than 'Does this face prove my dog knew?' That keeps the focus on safety, the environment and the next achievable change. You do not have to dismiss your dog's emotions to recognise that a familiar expression cannot reliably reconstruct an earlier event.

For a concise way to record your own observations, use the Dog Behavior Context Checklist.

Sources

[S1] Horowitz, A. (2009). Disambiguating the 'guilty look': salient prompts to a familiar dog behaviour. Behavioural Processes, 81(3), 447–452. DOI: 10.1016/j.beproc.2009.03.014.

[S2] Ostojić, L., Tkalčić, M., & Clayton, N. S. (2015). Are owners' reports of their dogs' 'guilty look' influenced by the dogs' action and evidence of the misdeed? Behavioural Processes, 111, 97–100. DOI: 10.1016/j.beproc.2014.12.010.