Dogs can eavesdrop on people’s reactions in third-party interactions.

By Claudia Bensimoun

First published: World Cynosport Rally| USDAA

Claudia Bensimoun served as the Veterinary Research Writer for World Cynosport Rally | USDAA from 2013 to 2020, where she specialized in translating canine sports and behavior research for the public. She now applies that expertise as a dog website content and SEO specialist.

Analysis of a University of Buenos Aires study revealing how dogs observe human interactions to choose who to approach. Posted Date: March 13, 2014, by Claudia Bensimoun.

What if your dog’s decisions are based on more than just direct commands? Groundbreaking research from the University of Buenos Aires, led by Dr. Bentosela’s team, provides compelling evidence that dogs are sophisticated social observers.

The study, “Can Dogs Learn from Eavesdropping?”, demonstrates that dogs use the outcomes of third-party interactions to form preferences and guide their own behavior. This analysis breaks down the experimental findings to explain how your dog may be learning simply by watching you.

Have you wondered how our reactions towards other people may affect the way our dogs react to the very same people? Can our body language affect the way our dogs behave?

Past research has demonstrated that the socio-cognitive abilities of dogs (in particular, dogs that can read human communicative gestures and cues) may be the result of the domestication process. This research has shown that dogs can read human emotions and determine whether an approaching person is friendly or not. It has also been demonstrated that dogs tend to prefer people who give them social rewards such as petting and positive verbalizations.

Dogs have been known to differentiate between a smiling face and a neutral face, as well as between expressions of happiness and disgust, and they also use this information to find food. Dogs are also able to recognize sadness and will approach a person who is crying. With that said, researchers expected dogs to be good candidates for a new study.

Research by Drs. Freidin, Putrino, D’Orazio, and Dr. Bentosela at the University of Buenos Aires, Argentina, demonstrated that dogs’ nonrandom choices of which person (called a “donor”) to approach for food relied on the simultaneous presence of multiple clues, such as the place where donors stood and several other features such as the behavior of someone requesting food (called the “beggar”), which included gestural and verbal reactions, and eating behaviors.

In the first experiment, 1a, the researchers assessed whether the dogs could develop a preference between the donors who behaved similarly and beggars asking for food. The donors would either react positively or negatively, using hand and body movements, as well as verbal means.

If the dogs developed a preference, it would mean that they were capable of discriminating between the beggar’s positive and negative emotional reactions, and that they were associating those specific reactions with the corresponding donor. It also meant that these dogs were using the information so that they would know which donor to go to (depending on whether the donor was friendly or not to the beggar).

In the second and third experiments, 1b and 1c, the researchers tested whether dogs were conditioned to the place, instead of to the donors. In Experiment 1b, the donors switched places between demonstrations before the dog could choose. In Experiment 1c, the phantom control group, the beggar had to present the same verbal and gestural cues as those shown in Experiment 1a, without donors and without the social interactive component.

Experiment

Seventy-two domestic dogs were recruited. The average age of these dogs was 4.73 years. Forty-one of these dogs were male and 31 were female. There were 17 Poodles, 5 German Shepherds, 5 Labrador Retrievers, 3 Golden Retrievers, 2 Cocker Spaniels, 1 Beagle, 1 Border Collie, 1 Boxer, 1 Breton, 1 Dalmatian, 1 Fox Terrier, 1 French Bulldog, 1 Great Dane, 1 Pitt Bull Terrier, 1 Samoyed, 1 Shih Tzu, 1 Weimaraner, 1 Yorkshire, and 27 mixed breeds. Of all the dogs, 36 had previous experience in other communicative tasks.

Subjects

All the subjects were randomly assigned to one of the three groups:

• Group VG (Verbal and Gestural): The beggar used both a verbal cue (“Can I have some?”) and a gesture (extending an arm with the palm up toward the donor).
• Group G (Gestural Only): The beggar used only the silent arm/hand gesture.
• Group V (Verbal Only): The beggar used only the verbal cue, keeping their hands behind their back.

This design allowed the researchers to isolate which type of human communication was most influential in the dogs’ eavesdropping.

Materials

All the dogs were tested in a familiar environment, either in their home or at a dog care facility that they sometimes attended. During the 5-10 minutes that it took the researchers to set up the experiment with the camera and tape, the dogs were allowed to interact with the assistant and the owner to become comfortable.

The key materials included:

• Two “Donors”: Actors who stood behind small tables, each with a piece of food visible.
• A “Beggar”: An actor who would approach the donors to request food.
• Food Rewards: Used to reinforce the donors’ roles (one generous, one selfish).
• A Camera: To record the dogs’ choices and behaviors for accurate analysis.
• Measuring Tape: To ensure consistent distances between all participants (donors, beggars, and the dog’s starting point).

Procedure & Key Results

The procedure involved an “observation phase” where the dog watched the beggar interact with each donor twice. One donor was “Generous” (giving food with a positive voice and smile), and the other was “Selfish” (refusing food with a negative voice and frown). After this, in the “test phase,” both donors called the dog neutrally. The dog’s choice to approach one donor was the critical data point.

The results were clear and significant:

1. Dogs Can Eavesdrop: In the main experiment (Group VG), dogs significantly preferred to approach the Generous donor over the Selfish one. They used the information from the third-party interaction to guide their own social choices.
2. Gestures are Key: The results varied by group. Dogs in the Gestural & Verbal (VG) and Gestural Only (G) groups successfully chose the generous donor. However, dogs in the Verbal Only (V) group did not show a significant preference. This indicates that visual cues (body language, gestures) were more critical for the dogs than vocal tone alone in this eavesdropping context.
3. They Follow the Person, Not the Place: When the donors switched places after the demonstration, the dogs still followed the generous person, not the location. This proved their choice was based on social assessment, not a simple associative cue.
4. The Social Interaction is Crucial: A control group confirmed that without the actual social interaction between the beggar and donors (just the beggar performing the cues alone), dogs did not develop a preference, ruling out that they were reacting to the beggar’s performance.

This study provides robust evidence that dogs are sophisticated social eavesdroppers. They don’t just learn from direct experience but can observe and evaluate human interactions to form judgments about who is likely to be cooperative. This ability relies heavily on reading human body language and gestures, underscoring the depth of the cross-species bond we share with our canine companions.

The groundbreaking 2013 study, ‘Dogs’ Eavesdropping from People’s Reactions in Third Party Interactions’ (Freidin et al.), published in PLoS ONE, provides the definitive evidence for this behavior

canine-eavesdropping, third-party-interaction, social-cognition-dogs, Freidin-study-2013, Bentosela-lab-research, University-Buenos-Aires,dog-social-learning, canine-gesture-comprehension, dog-verbal-cues, canine-decision-making, dog-social-evaluation, human-dog-communication, Claudia-Bensimoun-analysis, research-breakdown-expert, PLoS-One-study, evidence-based-canine-care, canine-cognition-research

© [2013-2025] BarkUpToday. All Rights Reserved.