Written by Dogoos Editorial · Editorial team

You walk through the door, spot the shredded cushion or the puddle by the rug, and there it is, the lowered head, the flattened ears, the slow tail and those wide white-rimmed eyes that seem to say sorry. It is one of the most convincing performances in the animal world, and it has fooled loving owners for generations. The truth is gentler and stranger than guilt, and once you know what that face really tracks, your whole relationship with everyday accidents changes for the better.

What you are looking at is not a confession, it is a negotiation, and your dog is doing their level best to keep the peace. The study everyone quotes is small and beautifully simple: the psychologist Alexandra Horowitz filmed fourteen pet dogs told to leave a tempting treat while their owner stepped out of the room, then told the owners on the way back in, sometimes truthfully and sometimes not, whether the dog had eaten it.

The guilty look turned up just as readily in the dogs that had obeyed, because what decided it was the owner and not the deed: the same dogs produced far more appeasing behaviour when they were told off than when they got a friendly hello. Stranger still, scolding pulled the strongest response out of the dogs that had been obedient, so the guiltiest-looking animal in the room was often the innocent one being told off for nothing.

Horowitz reads the display as a ritualised submissive act, stimulated by the learned expectation of punishment, and notes that your voice alone, a name said in a rising and accusatory tone, is often enough to bring it out. Fourteen dogs is a small sample, so a second team came at the question from the other end and asked whether the look is any use as information.

They let some dogs eat a forbidden food while the owner was out and quietly removed it from the rest, then asked ninety-six owners to judge, purely from the greeting they got at the door, whether their dog had eaten it. The owners were right no more often than chance, and the greetings were unmoved both by what the dog had done and by whether the evidence was still lying there in plain view. The look is a real signal, then, just not a signal about the crime. So what does a telling-off at the door actually teach.

The RSPCA puts the mechanism plainly: the punishment gets linked with your return, rather than with the chewing or the toileting that happened some time before. Your dog starts to dread the next homecoming, and that anxiety, the charity warns, makes chewing and lost toilet control more likely rather than less. Then the loop closes, because a dog who has been told off on arrival comes back smaller, head low, ears flat, tail tucked, and the owner reads that as guilt and tells them off again.

Whether tense homecomings leave a lasting mark has not, as far as we can find, been measured directly, but the nearest evidence gives pause: of ninety-two pet dogs, the ones taught with aversive methods showed more stress behaviour, bigger rises in cortisol and a gloomier showing on a cognitive bias test than the ones taught with rewards, inside the training ground and out of it. With that in mind, here are the three responses to retire for good.

The first is the delayed telling-off, which attaches itself to your arrival rather than to the act it is meant to be about. The second is walking your dog over to the mess and making them look at what they have done: in a survey of a hundred and forty owners at an American veterinary behaviour clinic, ten had tried rubbing the nose of a soiling dog in the patch, and not one of them reported that it helped; eight said it changed nothing at all and one said it made things worse.

Aggression tracked the confrontational methods right across that survey, with hitting or kicking drawing an aggressive response from forty-three per cent of the dogs it was tried on, in a group already referred for behaviour problems rather than the average dog on your street. The third, and the quiet root of the other two, is treating that appeasing face as proof your dog knows what they did.

Horowitz is careful here in a way the internet is not, because her findings do not show that dogs cannot feel guilt, only that what a behaviour study can measure is what an animal does and when. What her work rules out is the face as evidence, since it follows your tone rather than their memory. So what should you do instead, on a real Tuesday with a real dog.

Start by arranging the house so the mistake cannot happen, which is what the American Veterinary Society of Animal Behavior advises: manage the environment, avoid the situations that set the behaviour up, reinforce the thing you want instead, and leave out anything that works through force, pain or discomfort. If shoes get chewed, the shoes go in a closed cupboard and a good chew toy goes on the floor in their place; if the bin gets raided, it moves behind a cabinet door, and a baby gate keeps an unsupervised dog out of trouble while the habit builds.

Reward the behaviour you want the moment it appears, so chewing the right toy, settling on the mat or toileting in the right spot earns warm praise and a small treat right then. And when you catch a mishap unfolding in real time, interrupt it calmly with a cheerful noise or a swap for something allowed, then reward the switch, no anger required.

As for how long it takes, nobody can honestly hand you a number, so treat any article that promises you a tidy fortnight as guessing, expect boring practice rather than one dramatic lesson, and count the odd slip as learning rather than defiance. The mindset shift is this, you are not a judge handing down verdicts, you are a coach setting your dog up to win, and every accident is feedback about the environment rather than a flaw in your dog.

Picture the next time you come home to a mess: take a breath, clear it up without a word, and ask yourself what one change would stop it happening tomorrow. Your dog was never trying to upset you, they were trying to stay close to you, and when you meet that with calm and a plan the guilty look fades and an easy welcome takes its place. One thing not to file under training: toileting indoors from an adult dog who used to be reliable, or that starts abruptly, is a medical question first.

A veterinary behaviourist writing for vets puts medical causes at the top of the list in exactly those cases and says every house-soiling dog should be examined, so make the appointment before you build the training plan. You have got this, and so does your dog.

Sources & further reading

1. Disambiguating the "guilty look": Salient prompts to a familiar dog behaviour — Behavioural Processes (Elsevier) 81(3):447–452, DOI 10.1016/j.beproc.2009.03.014 (2009-07-01)
2. Are owners' reports of their dogs' 'guilty look' influenced by the dogs' action and evidence of the misdeed? — Behavioural Processes (Elsevier) 111:97–100, DOI 10.1016/j.beproc.2014.12.010 (2015-02-01)
3. Training your dog to be left alone — RSPCA (rspca.org.uk)
4. Position Statement on Humane Dog Training — American Veterinary Society of Animal Behavior (AVSAB), © 2021
5. Does training method matter? Evidence for the negative impact of aversive-based methods on companion dog welfare — PLOS ONE 15(12):e0225023, DOI 10.1371/journal.pone.0225023 (2020-12-16)
6. Survey of the use and outcome of confrontational and non-confrontational training methods in client-owned dogs showing undesired behaviors — Applied Animal Behaviour Science (Elsevier) 117:47–54, DOI (2009-02-01)
7. Canine House Soiling: Back to Basics — Today's Veterinary Practice (Nov/Dec 2017 issue) (2017-10-06)